All update up to proxy

AI.PROMPT

@@ -383,7 +383,9 @@ Kiro Gateway is a third-party proxy gateway that provides OpenAI and Anthropic-c
 
 **Configuration:**
 
-In `config/providers.json`:
+**IMPORTANT:** Kiro providers use a different configuration structure than other providers. Instead of an `api_key` field, they use a `kiro_config` object with authentication settings specific to Kiro/Amazon Q Developer.
+
+**Standard Provider Configuration (Google, OpenAI, Anthropic, etc.):**
 ```json
 {
   "gemini": {
@@ -405,18 +407,32 @@ In `config/providers.json`:
         "context_size": 1000000
       }
     ]
-  },
+  }
+}
+```
+
+**Kiro Provider Configuration (uses kiro_config instead of api_key):**
+```json
+{
   "kiro": {
     "id": "kiro",
-    "name": "Kiro Gateway (Amazon Q Developer)",
-    "endpoint": "http://localhost:8000/v1",
+    "name": "Kiro (Amazon Q Developer)",
+    "endpoint": "https://q.us-east-1.amazonaws.com",
     "type": "kiro",
-    "api_key_required": true,
-    "api_key": "YOUR_KIRO_API_KEY",
+    "api_key_required": false,
     "rate_limit": 0,
+    "kiro_config": {
+      "region": "us-east-1",
+      "creds_file": "~/.config/Code/User/globalStorage/amazon.q/credentials.json",
+      "sqlite_db": "",
+      "refresh_token": "",
+      "profile_arn": "",
+      "client_id": "",
+      "client_secret": ""
+    },
     "models": [
       {
-        "name": "claude-sonnet-4-5",
+        "name": "anthropic.claude-3-5-sonnet-20241022-v2:0",
         "rate_limit": 0,
         "max_request_tokens": 200000,
         "context_size": 200000
@@ -426,6 +442,20 @@ In `config/providers.json`:
 }
 ```
 
+**Kiro Configuration Fields (kiro_config):**
+- `region`: AWS region (default: "us-east-1")
+- `creds_file`: Path to Kiro IDE credentials JSON file (Option 1)
+- `sqlite_db`: Path to kiro-cli SQLite database (Option 2)
+- `refresh_token`: Direct refresh token for authentication (Option 3)
+- `profile_arn`: AWS CodeWhisperer profile ARN (optional)
+- `client_id`: OAuth client ID for AWS SSO OIDC (used with refresh_token)
+- `client_secret`: OAuth client secret for AWS SSO OIDC (used with refresh_token)
+
+Choose one authentication method:
+1. **Kiro IDE credentials**: Set `creds_file` to the path of your Kiro IDE credentials
+2. **kiro-cli database**: Set `sqlite_db` to the path of your kiro-cli SQLite database
+3. **Direct credentials**: Set `refresh_token` and optionally `client_id`/`client_secret` for AWS SSO OIDC
+
 In `config/rotations.json` (API keys are now referenced from providers.json):
 ```json
 {
@@ -590,6 +620,25 @@ This AI.PROMPT file is automatically updated when significant changes are made t
 
 ### Recent Updates
 
+**2026-03-23 - Proxy-Awareness Implementation**
+- Added ProxyHeadersMiddleware to handle reverse proxy deployments
+- Implemented automatic detection of proxy headers (X-Forwarded-Proto, X-Forwarded-Host, X-Forwarded-Port, X-Forwarded-Prefix, X-Forwarded-For)
+- Updated all RedirectResponse calls to use proxy-aware url_for() function
+- Updated templates to use proxy-aware URL generation
+- Added comprehensive nginx proxy configuration examples to DOCUMENTATION.md
+- Application now fully supports deployment behind reverse proxies with subpaths
+- Dashboard URLs, redirects, and API responses automatically adjust based on proxy configuration
+
+**2026-03-23 - Kiro Provider Dashboard Support**
+- Updated providers.html dashboard to show kiro-specific configuration fields
+- Kiro providers now display `kiro_config` options (region, creds_file, sqlite_db, refresh_token, profile_arn, client_id, client_secret) instead of standard api_key fields
+- Added `updateProviderType()` function to handle switching between kiro and other provider types
+- Added `updateKiroConfig()` function to manage kiro-specific configuration fields
+- Updated `addProvider()` function to ask for provider type and set appropriate defaults for kiro providers
+- Dashboard automatically initializes kiro_config when provider type is changed to "kiro"
+- Dashboard automatically removes kiro_config when provider type is changed from "kiro" to another type
+- Updated AI.PROMPT documentation to clarify kiro provider configuration differences
+
 **2026-03-22 - Configuration Refactoring**
 - Centralized API key storage in providers.json
 - API keys are now stored only in provider definitions, not in rotation/autoselect configs
@@ -842,3 +891,117 @@ When making changes that affect the project structure or functionality:
 ### Version Control
 
 AI.PROMPT should be committed to version control along with other project files to ensure that documentation stays synchronized with the codebase.
+
+## Proxy-Awareness and Reverse Proxy Support
+
+### Overview
+
+AISBF is fully proxy-aware and can be deployed behind reverse proxies like nginx, Apache, or Caddy. The application automatically detects and respects standard proxy headers to ensure correct URL generation and routing.
+
+### Implementation
+
+**Proxy Headers Middleware** ([`main.py`](main.py:349-408)):
+- `ProxyHeadersMiddleware` class handles proxy header detection
+- Automatically updates request scope with proxy information
+- Processes standard proxy headers before other middleware
+- Must be added before other middleware to ensure proper request scope modification
+
+**Supported Proxy Headers**:
+- `X-Forwarded-Proto`: Original protocol (http/https)
+- `X-Forwarded-Host`: Original hostname
+- `X-Forwarded-Port`: Original port number
+- `X-Forwarded-Prefix` or `X-Script-Name`: URL subpath/prefix
+- `X-Forwarded-For`: Original client IP address
+
+**Helper Functions** ([`main.py`](main.py:410-445)):
+- `get_base_url(request)`: Returns full base URL respecting proxy configuration
+- `url_for(request, path)`: Generates proxy-aware URLs for any path
+- `setup_template_globals()`: Configures Jinja2 templates with proxy-aware URL functions
+
+### URL Generation
+
+**In Python Code**:
+```python
+# Use url_for() for all redirects
+return RedirectResponse(url=url_for(request, "/dashboard"), status_code=303)
+
+# Get base URL
+base_url = get_base_url(request)  # e.g., "https://example.com:8443/aisbf"
+```
+
+**In Jinja2 Templates**:
+```html
+<!-- Use url_for() function in templates -->
+<a href="{{ url_for(request, '/dashboard/providers') }}">Providers</a>
+
+<!-- For JavaScript fetch calls -->
+fetch('{{ url_for(request, "/dashboard/restart") }}', {
+    method: 'POST',
+    headers: {'Content-Type': 'application/json'}
+})
+```
+
+### Deployment Scenarios
+
+**Scenario 1: Root Path Deployment**
+- URL: `https://aisbf.example.com/`
+- No special configuration needed
+- Proxy sets: `X-Forwarded-Proto: https`, `X-Forwarded-Host: aisbf.example.com`
+
+**Scenario 2: Subpath Deployment**
+- URL: `https://example.com/aisbf/`
+- Proxy must set: `X-Forwarded-Prefix: /aisbf`
+- AISBF automatically adjusts all URLs to include `/aisbf` prefix
+
+**Scenario 3: Custom Port**
+- URL: `https://example.com:8443/`
+- Proxy sets: `X-Forwarded-Port: 8443`
+- URLs include port number when non-standard
+
+### Nginx Configuration
+
+See [`DOCUMENTATION.md`](DOCUMENTATION.md) "Proxy Deployment" section for complete nginx configuration examples including:
+- Simple reverse proxy (root path)
+- Reverse proxy with subpath
+- Load balancing multiple instances
+- API-only deployment with rate limiting
+
+### Testing Proxy Configuration
+
+After deploying behind a proxy:
+1. Access the dashboard and verify all links work correctly
+2. Test redirects (login, logout) work properly
+3. Verify the restart button uses correct URL
+4. Test API endpoints through the proxy
+5. Check that streaming responses work correctly
+
+### Common Issues and Solutions
+
+**Issue**: Dashboard links return 404 errors
+**Solution**: Ensure `X-Forwarded-Prefix` header is set correctly in nginx
+
+**Issue**: Redirect loops after login
+**Solution**: Verify `X-Forwarded-Proto` is set to `https` for HTTPS deployments
+
+**Issue**: URLs missing subpath prefix
+**Solution**: Check that `X-Forwarded-Prefix` or `X-Script-Name` header is configured
+
+**Issue**: Streaming responses not working
+**Solution**: Ensure `proxy_buffering off` is set in nginx configuration
+
+### Files Modified for Proxy-Awareness
+
+- [`main.py`](main.py): Added ProxyHeadersMiddleware, url_for(), get_base_url() functions
+- [`templates/base.html`](templates/base.html): Updated all URLs to use url_for() function
+- All dashboard route handlers in [`main.py`](main.py): Updated RedirectResponse calls to use url_for()
+- [`DOCUMENTATION.md`](DOCUMENTATION.md): Added comprehensive nginx proxy configuration examples
+
+### Security Considerations
+
+When deploying behind a proxy:
+- AISBF should bind to localhost only (default: `127.0.0.1:17765`)
+- Proxy should handle SSL/TLS termination
+- Enable authentication in AISBF configuration
+- Configure rate limiting in the proxy
+- Set security headers in proxy configuration
+- Monitor access logs for suspicious activity
\ No newline at end of file

API_EXAMPLES.md

@@ -3,23 +3,52 @@
 This document provides practical examples for using the AISBF API endpoints.
 
 ## Table of Contents
-- [OpenAI-Compatible v1 Endpoints](#openai-compatible-v1-endpoints)
+- [Three Proxy Paths](#three-proxy-paths)
 - [Chat Completions](#chat-completions)
 - [Audio Endpoints](#audio-endpoints)
 - [Image Generation](#image-generation)
 - [Embeddings](#embeddings)
 - [Model Listing](#model-listing)
-- [Rotations](#rotations)
-- [Autoselect](#autoselect)
+- [Advanced Features](#advanced-features)
+
+## Three Proxy Paths
+
+AISBF provides three ways to proxy AI models:
+
+### PATH 1: Direct Provider Models
+Format: `{provider_id}/{model_name}`
+```bash
+# Examples:
+"openai/gpt-4"
+"gemini/gemini-2.0-flash"
+"anthropic/claude-3-5-sonnet-20241022"
+"kilotest/kilo/free"
+```
+
+### PATH 2: Rotations
+Format: `rotation/{rotation_name}`
+```bash
+# Examples:
+"rotation/coding"
+"rotation/general"
+```
+
+### PATH 3: Autoselect
+Format: `autoselect/{autoselect_name}`
+```bash
+# Examples:
+"autoselect/autoselect"
+```
 
 ## OpenAI-Compatible v1 Endpoints
 
-The v1 endpoints follow the standard OpenAI API format with `provider/model` notation.
+The v1 endpoints follow the standard OpenAI API format and support all three proxy paths.
 
 ### Chat Completions
 
-#### Using cURL
+#### PATH 1: Direct Provider Models
 
+Using cURL:
 ```bash
 curl -X POST http://localhost:17765/api/v1/chat/completions \
   -H "Content-Type: application/json" \
@@ -31,38 +60,120 @@ curl -X POST http://localhost:17765/api/v1/chat/completions \
   }'
 ```
 
-#### Using Python
-
+Using Python with OpenAI SDK:
 ```python
-import requests
+from openai import OpenAI
 
-response = requests.post(
-    "http://localhost:17765/api/v1/chat/completions",
-    json={
-        "model": "openai/gpt-4",
-        "messages": [
+client = OpenAI(
+    base_url="http://localhost:17765/api/v1",
+    api_key="dummy"  # Not required if auth is disabled
+)
+
+response = client.chat.completions.create(
+    model="openai/gpt-4",
+    messages=[
         {"role": "user", "content": "Hello, how are you?"}
     ]
-    }
 )
 
-print(response.json())
+print(response.choices[0].message.content)
+```
+
+Different providers:
+```bash
+# Google Gemini
+curl -X POST http://localhost:17765/api/v1/chat/completions \
+  -H "Content-Type: application/json" \
+  -d '{
+    "model": "gemini/gemini-2.0-flash",
+    "messages": [{"role": "user", "content": "Hello"}]
+  }'
+
+# Anthropic Claude
+curl -X POST http://localhost:17765/api/v1/chat/completions \
+  -H "Content-Type: application/json" \
+  -d '{
+    "model": "anthropic/claude-3-5-sonnet-20241022",
+    "messages": [{"role": "user", "content": "Hello"}]
+  }'
+
+# Custom provider with nested model path
+curl -X POST http://localhost:17765/api/v1/chat/completions \
+  -H "Content-Type: application/json" \
+  -d '{
+    "model": "kilotest/kilo/free",
+    "messages": [{"role": "user", "content": "Hello"}]
+  }'
+
+# Ollama (local)
+curl -X POST http://localhost:17765/api/v1/chat/completions \
+  -H "Content-Type: application/json" \
+  -d '{
+    "model": "ollama/llama2",
+    "messages": [{"role": "user", "content": "Hello"}]
+  }'
 ```
 
-#### Using Python with OpenAI SDK
+#### PATH 2: Rotations (Load Balancing)
 
+Using cURL:
+```bash
+curl -X POST http://localhost:17765/api/v1/chat/completions \
+  -H "Content-Type: application/json" \
+  -d '{
+    "model": "rotation/coding",
+    "messages": [
+      {"role": "user", "content": "Write a Python function to sort a list"}
+    ]
+  }'
+```
+
+Using Python with OpenAI SDK:
 ```python
 from openai import OpenAI
 
 client = OpenAI(
     base_url="http://localhost:17765/api/v1",
-    api_key="dummy"  # Not required if auth is disabled
+    api_key="dummy"
 )
 
 response = client.chat.completions.create(
-    model="openai/gpt-4",
+    model="rotation/coding",
     messages=[
-        {"role": "user", "content": "Hello, how are you?"}
+        {"role": "user", "content": "Write a Python function to sort a list"}
+    ]
+)
+
+print(response.choices[0].message.content)
+```
+
+#### PATH 3: Autoselect (AI-Powered Selection)
+
+Using cURL:
+```bash
+curl -X POST http://localhost:17765/api/v1/chat/completions \
+  -H "Content-Type: application/json" \
+  -d '{
+    "model": "autoselect/autoselect",
+    "messages": [
+      {"role": "user", "content": "Debug this Python code: def add(a,b): return a-b"}
+    ]
+  }'
+```
+
+Using Python with OpenAI SDK:
+```python
+from openai import OpenAI
+
+client = OpenAI(
+    base_url="http://localhost:17765/api/v1",
+    api_key="dummy"
+)
+
+response = client.chat.completions.create(
+    model="autoselect/autoselect",
+    messages=[
+        {"role": "user", "content": "Debug this Python code: def add(a,b): return a-b"}
     ]
 )
 
@@ -71,6 +182,7 @@ print(response.choices[0].message.content)
 
 #### Streaming Response
 
+Works with all three proxy paths:
 ```python
 from openai import OpenAI
 
@@ -79,11 +191,24 @@ client = OpenAI(
     api_key="dummy"
 )
 
+# PATH 1: Direct provider
 stream = client.chat.completions.create(
     model="gemini/gemini-2.0-flash",
-    messages=[
-        {"role": "user", "content": "Write a short poem"}
-    ],
+    messages=[{"role": "user", "content": "Write a short poem"}],
+    stream=True
+)
+
+# PATH 2: Rotation
+stream = client.chat.completions.create(
+    model="rotation/coding",
+    messages=[{"role": "user", "content": "Write a short poem"}],
+    stream=True
+)
+
+# PATH 3: Autoselect
+stream = client.chat.completions.create(
+    model="autoselect/autoselect",
+    messages=[{"role": "user", "content": "Write a short poem"}],
     stream=True
 )
 
@@ -92,44 +217,27 @@ for chunk in stream:
         print(chunk.choices[0].delta.content, end="")
 ```
 
-#### Using Different Providers
+## Audio Endpoints
 
-```bash
-# Google Gemini
-curl -X POST http://localhost:17765/api/v1/chat/completions \
-  -H "Content-Type: application/json" \
-  -d '{
-    "model": "gemini/gemini-2.0-flash",
-    "messages": [{"role": "user", "content": "Hello"}]
-  }'
+**Note:** Audio endpoints support all three proxy paths (direct providers, rotations, and autoselect).
 
-# Anthropic Claude
-curl -X POST http://localhost:17765/api/v1/chat/completions \
-  -H "Content-Type: application/json" \
-  -d '{
-    "model": "anthropic/claude-3-5-sonnet-20241022",
-    "messages": [{"role": "user", "content": "Hello"}]
-  }'
+### Audio Transcription
 
-# Ollama (local)
-curl -X POST http://localhost:17765/api/v1/chat/completions \
-  -H "Content-Type: application/json" \
-  -d '{
-    "model": "ollama/llama2",
-    "messages": [{"role": "user", "content": "Hello"}]
-  }'
+Using `/api/audio/transcriptions`:
+```bash
+curl -X POST http://localhost:17765/api/audio/transcriptions \
+  -F "file=@audio.mp3" \
+  -F "model=openai/whisper-1"
 ```
 
-## Audio Endpoints
-
-### Audio Transcription
-
+Using `/api/v1/audio/transcriptions` (OpenAI-compatible):
 ```bash
 curl -X POST http://localhost:17765/api/v1/audio/transcriptions \
   -F "file=@audio.mp3" \
   -F "model=openai/whisper-1"
 ```
 
+Using Python with OpenAI SDK:
 ```python
 from openai import OpenAI
 
@@ -149,8 +257,9 @@ print(transcript.text)
 
 ### Text-to-Speech
 
+Using `/api/audio/speech`:
 ```bash
-curl -X POST http://localhost:17765/api/v1/audio/speech \
+curl -X POST http://localhost:17765/api/audio/speech \
   -H "Content-Type: application/json" \
   -d '{
     "model": "openai/tts-1",
@@ -160,6 +269,7 @@ curl -X POST http://localhost:17765/api/v1/audio/speech \
   --output speech.mp3
 ```
 
+Using Python with OpenAI SDK:
 ```python
 from openai import OpenAI
 
@@ -179,8 +289,11 @@ response.stream_to_file("speech.mp3")
 
 ## Image Generation
 
+**Note:** Image generation supports all three proxy paths (direct providers, rotations, and autoselect).
+
+Using `/api/images/generations`:
 ```bash
-curl -X POST http://localhost:17765/api/v1/images/generations \
+curl -X POST http://localhost:17765/api/images/generations \
   -H "Content-Type: application/json" \
   -d '{
     "model": "openai/dall-e-3",
@@ -190,6 +303,7 @@ curl -X POST http://localhost:17765/api/v1/images/generations \
   }'
 ```
 
+Using Python with OpenAI SDK:
 ```python
 from openai import OpenAI
 
@@ -210,8 +324,11 @@ print(response.data[0].url)
 
 ## Embeddings
 
+**Note:** Embeddings support all three proxy paths (direct providers, rotations, and autoselect).
+
+Using `/api/embeddings`:
 ```bash
-curl -X POST http://localhost:17765/api/v1/embeddings \
+curl -X POST http://localhost:17765/api/embeddings \
   -H "Content-Type: application/json" \
   -d '{
     "model": "openai/text-embedding-ada-002",
@@ -219,6 +336,7 @@ curl -X POST http://localhost:17765/api/v1/embeddings \
   }'
 ```
 
+Using Python with OpenAI SDK:
 ```python
 from openai import OpenAI
 
@@ -237,12 +355,27 @@ print(response.data[0].embedding)
 
 ## Model Listing
 
-### List All Models
+### List All Models (All Three Proxy Paths)
 
+The `/api/models` endpoint lists models from all three proxy paths:
+
+Using cURL:
 ```bash
-curl http://localhost:17765/api/v1/models
+curl http://localhost:17765/api/models
+```
+
+Using Python:
+```python
+import requests
+
+response = requests.get("http://localhost:17765/api/models")
+models = response.json()["data"]
+
+for model in models:
+    print(f"{model['id']} - Type: {model.get('type', 'unknown')}")
 ```
 
+Using OpenAI SDK:
 ```python
 from openai import OpenAI
 
@@ -256,99 +389,59 @@ for model in models.data:
     print(f"{model.id} - {model.owned_by}")
 ```
 
-## Rotations
+Example output:
+```
+openai/gpt-4 - Type: provider
+gemini/gemini-2.0-flash - Type: provider
+rotation/coding - Type: rotation
+rotation/general - Type: rotation
+autoselect/autoselect - Type: autoselect
+```
 
-Rotations provide weighted load balancing across multiple providers.
+## Legacy Endpoints
 
-### Using Rotation with v1 Endpoint
+For backward compatibility, these endpoints are still available:
 
+### Legacy Provider Endpoints
 ```bash
-curl -X POST http://localhost:17765/api/v1/chat/completions \
+# Direct provider access (model without provider prefix)
+curl -X POST http://localhost:17765/api/openai/chat/completions \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "coding",
-    "messages": [
-      {"role": "user", "content": "Write a Python function to sort a list"}
-    ]
+    "model": "gpt-4",
+    "messages": [{"role": "user", "content": "Hello"}]
   }'
-```
 
-### Using Legacy Rotation Endpoint
-
-```bash
-curl -X POST http://localhost:17765/api/rotations/chat/completions \
-  -H "Content-Type: application/json" \
-  -d '{
-    "model": "coding",
-    "messages": [
-      {"role": "user", "content": "Write a Python function to sort a list"}
-    ]
-  }'
+# List provider models
+curl http://localhost:17765/api/openai/models
 ```
 
-### List Available Rotations
-
+### Legacy Rotation Endpoints
 ```bash
+# List rotations
 curl http://localhost:17765/api/rotations
-```
 
-### List Rotation Models
-
-```bash
-curl http://localhost:17765/api/rotations/models
-```
-
-## Autoselect
-
-Autoselect uses AI to automatically select the best model based on your request.
-
-### Using Autoselect with v1 Endpoint
-
-```bash
-curl -X POST http://localhost:17765/api/v1/chat/completions \
-  -H "Content-Type: application/json" \
-  -d '{
-    "model": "autoselect",
-    "messages": [
-      {"role": "user", "content": "Debug this Python code: def add(a,b): return a-b"}
-    ]
-  }'
-```
-
-### Using Legacy Autoselect Endpoint
-
-```bash
-curl -X POST http://localhost:17765/api/autoselect/chat/completions \
+# Use rotation (model name = rotation name)
+curl -X POST http://localhost:17765/api/rotations/chat/completions \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "autoselect",
-    "messages": [
-      {"role": "user", "content": "Debug this Python code: def add(a,b): return a-b"}
-    ]
+    "model": "coding",
+    "messages": [{"role": "user", "content": "Write code"}]
   }'
 ```
 
-### List Available Autoselect Configurations
-
+### Legacy Autoselect Endpoints
 ```bash
+# List autoselect configurations
 curl http://localhost:17765/api/autoselect
-```
-
-## Legacy Provider Endpoints
 
-You can also use provider-specific endpoints:
-
-```bash
-# Direct provider access
-curl -X POST http://localhost:17765/api/openai/chat/completions \
+# Use autoselect (model name = autoselect name)
+curl -X POST http://localhost:17765/api/autoselect/chat/completions \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "gpt-4",
-    "messages": [{"role": "user", "content": "Hello"}]
+    "model": "autoselect",
+    "messages": [{"role": "user", "content": "Help me"}]
   }'
-
-# List provider models
-curl http://localhost:17765/api/openai/models
 ```
 
 ## Authentication
@@ -383,6 +476,7 @@ response = client.chat.completions.create(
 
 ### Using fetch API
 
+PATH 1: Direct Provider
 ```javascript
 const response = await fetch('http://localhost:17765/api/v1/chat/completions', {
   method: 'POST',
@@ -401,8 +495,47 @@ const data = await response.json();
 console.log(data.choices[0].message.content);
 ```
 
+PATH 2: Rotation
+```javascript
+const response = await fetch('http://localhost:17765/api/v1/chat/completions', {
+  method: 'POST',
+  headers: {
+    'Content-Type': 'application/json',
+  },
+  body: JSON.stringify({
+    model: 'rotation/coding',
+    messages: [
+      { role: 'user', content: 'Write a sorting function' }
+    ]
+  })
+});
+
+const data = await response.json();
+console.log(data.choices[0].message.content);
+```
+
+PATH 3: Autoselect
+```javascript
+const response = await fetch('http://localhost:17765/api/v1/chat/completions', {
+  method: 'POST',
+  headers: {
+    'Content-Type': 'application/json',
+  },
+  body: JSON.stringify({
+    model: 'autoselect/autoselect',
+    messages: [
+      { role: 'user', content: 'Help me with this task' }
+    ]
+  })
+});
+
+const data = await response.json();
+console.log(data.choices[0].message.content);
+```
+
 ### Using OpenAI SDK
 
+Works with all three proxy paths:
 ```javascript
 import OpenAI from 'openai';
 
@@ -411,18 +544,30 @@ const client = new OpenAI({
   apiKey: 'dummy'
 });
 
-const response = await client.chat.completions.create({
+// PATH 1: Direct provider
+const response1 = await client.chat.completions.create({
   model: 'openai/gpt-4',
-  messages: [
-    { role: 'user', content: 'Hello, how are you?' }
-  ]
+  messages: [{ role: 'user', content: 'Hello' }]
+});
+
+// PATH 2: Rotation
+const response2 = await client.chat.completions.create({
+  model: 'rotation/coding',
+  messages: [{ role: 'user', content: 'Write code' }]
+});
+
+// PATH 3: Autoselect
+const response3 = await client.chat.completions.create({
+  model: 'autoselect/autoselect',
+  messages: [{ role: 'user', content: 'Help me' }]
 });
 
-console.log(response.choices[0].message.content);
+console.log(response1.choices[0].message.content);
 ```
 
 ### Streaming in JavaScript
 
+Works with all three proxy paths:
 ```javascript
 import OpenAI from 'openai';
 
@@ -431,8 +576,9 @@ const client = new OpenAI({
   apiKey: 'dummy'
 });
 
+// Use any of the three proxy paths
 const stream = await client.chat.completions.create({
-  model: 'gemini/gemini-2.0-flash',
+  model: 'gemini/gemini-2.0-flash',  // or 'rotation/coding' or 'autoselect/autoselect'
   messages: [
     { role: 'user', content: 'Write a short poem' }
   ],

DOCUMENTATION.md

@@ -485,6 +485,377 @@ aisbf stop
 ```
 Stops running daemon and removes PID file.
 
+## Proxy Deployment
+
+AISBF is fully proxy-aware and can be deployed behind reverse proxies like nginx, Apache, or Caddy. The application automatically detects and respects standard proxy headers to ensure correct URL generation and routing.
+
+### Proxy-Awareness Features
+
+AISBF automatically handles:
+- **X-Forwarded-Proto**: Detects original protocol (http/https)
+- **X-Forwarded-Host**: Detects original hostname
+- **X-Forwarded-Port**: Detects original port
+- **X-Forwarded-Prefix** or **X-Script-Name**: Detects URL subpath/prefix
+- **X-Forwarded-For**: Detects original client IP address
+
+All URLs in the dashboard, redirects, and API responses are automatically adjusted based on these headers.
+
+### Nginx Proxy Configuration Examples
+
+#### Example 1: Simple Reverse Proxy (Root Path)
+
+Deploy AISBF at the root of a domain (e.g., `https://aisbf.example.com/`):
+
+```nginx
+server {
+    listen 80;
+    server_name aisbf.example.com;
+    
+    # Redirect HTTP to HTTPS
+    return 301 https://$server_name$request_uri;
+}
+
+server {
+    listen 443 ssl http2;
+    server_name aisbf.example.com;
+    
+    # SSL configuration
+    ssl_certificate /etc/ssl/certs/aisbf.example.com.crt;
+    ssl_certificate_key /etc/ssl/private/aisbf.example.com.key;
+    ssl_protocols TLSv1.2 TLSv1.3;
+    ssl_ciphers HIGH:!aNULL:!MD5;
+    
+    # Proxy settings
+    location / {
+        proxy_pass http://127.0.0.1:17765;
+        proxy_http_version 1.1;
+        
+        # Proxy headers for AISBF proxy-awareness
+        proxy_set_header Host $host;
+        proxy_set_header X-Real-IP $remote_addr;
+        proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;
+        proxy_set_header X-Forwarded-Proto $scheme;
+        proxy_set_header X-Forwarded-Host $host;
+        proxy_set_header X-Forwarded-Port $server_port;
+        
+        # WebSocket support for streaming
+        proxy_set_header Upgrade $http_upgrade;
+        proxy_set_header Connection "upgrade";
+        
+        # Timeouts for long-running requests
+        proxy_connect_timeout 60s;
+        proxy_send_timeout 300s;
+        proxy_read_timeout 300s;
+        
+        # Buffering settings for streaming
+        proxy_buffering off;
+        proxy_cache off;
+    }
+    
+    # Security headers
+    add_header X-Frame-Options "SAMEORIGIN" always;
+    add_header X-Content-Type-Options "nosniff" always;
+    add_header X-XSS-Protection "1; mode=block" always;
+    
+    # Logging
+    access_log /var/log/nginx/aisbf_access.log;
+    error_log /var/log/nginx/aisbf_error.log;
+}
+```
+
+#### Example 2: Reverse Proxy with Subpath
+
+Deploy AISBF at a subpath (e.g., `https://example.com/aisbf/`):
+
+```nginx
+server {
+    listen 443 ssl http2;
+    server_name example.com;
+    
+    # SSL configuration
+    ssl_certificate /etc/ssl/certs/example.com.crt;
+    ssl_certificate_key /etc/ssl/private/example.com.key;
+    ssl_protocols TLSv1.2 TLSv1.3;
+    ssl_ciphers HIGH:!aNULL:!MD5;
+    
+    # AISBF at /aisbf subpath
+    location /aisbf/ {
+        # Remove /aisbf prefix before proxying
+        rewrite ^/aisbf/(.*) /$1 break;
+        
+        proxy_pass http://127.0.0.1:17765;
+        proxy_http_version 1.1;
+        
+        # Proxy headers for AISBF proxy-awareness
+        proxy_set_header Host $host;
+        proxy_set_header X-Real-IP $remote_addr;
+        proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;
+        proxy_set_header X-Forwarded-Proto $scheme;
+        proxy_set_header X-Forwarded-Host $host;
+        proxy_set_header X-Forwarded-Port $server_port;
+        
+        # IMPORTANT: Set the subpath prefix
+        proxy_set_header X-Forwarded-Prefix /aisbf;
+        
+        # WebSocket support for streaming
+        proxy_set_header Upgrade $http_upgrade;
+        proxy_set_header Connection "upgrade";
+        
+        # Timeouts for long-running requests
+        proxy_connect_timeout 60s;
+        proxy_send_timeout 300s;
+        proxy_read_timeout 300s;
+        
+        # Buffering settings for streaming
+        proxy_buffering off;
+        proxy_cache off;
+    }
+    
+    # Other locations for your main application
+    location / {
+        # Your main application configuration
+        root /var/www/html;
+        index index.html;
+    }
+    
+    # Security headers
+    add_header X-Frame-Options "SAMEORIGIN" always;
+    add_header X-Content-Type-Options "nosniff" always;
+    add_header X-XSS-Protection "1; mode=block" always;
+    
+    # Logging
+    access_log /var/log/nginx/example_access.log;
+    error_log /var/log/nginx/example_error.log;
+}
+```
+
+#### Example 3: Load Balancing Multiple AISBF Instances
+
+Deploy multiple AISBF instances for high availability:
+
+```nginx
+upstream aisbf_backend {
+    # Multiple AISBF instances
+    server 127.0.0.1:17765 max_fails=3 fail_timeout=30s;
+    server 127.0.0.1:17766 max_fails=3 fail_timeout=30s;
+    server 127.0.0.1:17767 max_fails=3 fail_timeout=30s;
+    
+    # Load balancing method
+    least_conn;  # Route to instance with fewest connections
+    
+    # Session persistence (optional, for dashboard sessions)
+    ip_hash;
+}
+
+server {
+    listen 443 ssl http2;
+    server_name aisbf.example.com;
+    
+    # SSL configuration
+    ssl_certificate /etc/ssl/certs/aisbf.example.com.crt;
+    ssl_certificate_key /etc/ssl/private/aisbf.example.com.key;
+    ssl_protocols TLSv1.2 TLSv1.3;
+    ssl_ciphers HIGH:!aNULL:!MD5;
+    
+    location / {
+        proxy_pass http://aisbf_backend;
+        proxy_http_version 1.1;
+        
+        # Proxy headers for AISBF proxy-awareness
+        proxy_set_header Host $host;
+        proxy_set_header X-Real-IP $remote_addr;
+        proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;
+        proxy_set_header X-Forwarded-Proto $scheme;
+        proxy_set_header X-Forwarded-Host $host;
+        proxy_set_header X-Forwarded-Port $server_port;
+        
+        # WebSocket support for streaming
+        proxy_set_header Upgrade $http_upgrade;
+        proxy_set_header Connection "upgrade";
+        
+        # Timeouts for long-running requests
+        proxy_connect_timeout 60s;
+        proxy_send_timeout 300s;
+        proxy_read_timeout 300s;
+        
+        # Buffering settings for streaming
+        proxy_buffering off;
+        proxy_cache off;
+        
+        # Health check
+        proxy_next_upstream error timeout http_502 http_503 http_504;
+    }
+    
+    # Health check endpoint
+    location /health {
+        proxy_pass http://aisbf_backend/health;
+        access_log off;
+    }
+    
+    # Security headers
+    add_header X-Frame-Options "SAMEORIGIN" always;
+    add_header X-Content-Type-Options "nosniff" always;
+    add_header X-XSS-Protection "1; mode=block" always;
+    
+    # Logging
+    access_log /var/log/nginx/aisbf_access.log;
+    error_log /var/log/nginx/aisbf_error.log;
+}
+```
+
+#### Example 4: API-Only Deployment with Rate Limiting
+
+Deploy AISBF as an API-only service with rate limiting:
+
+```nginx
+# Define rate limiting zones
+limit_req_zone $binary_remote_addr zone=aisbf_api:10m rate=10r/s;
+limit_req_zone $binary_remote_addr zone=aisbf_dashboard:10m rate=5r/s;
+
+server {
+    listen 443 ssl http2;
+    server_name api.example.com;
+    
+    # SSL configuration
+    ssl_certificate /etc/ssl/certs/api.example.com.crt;
+    ssl_certificate_key /etc/ssl/private/api.example.com.key;
+    ssl_protocols TLSv1.2 TLSv1.3;
+    ssl_ciphers HIGH:!aNULL:!MD5;
+    
+    # API endpoints with rate limiting
+    location /api/ {
+        # Apply rate limiting
+        limit_req zone=aisbf_api burst=20 nodelay;
+        
+        proxy_pass http://127.0.0.1:17765;
+        proxy_http_version 1.1;
+        
+        # Proxy headers for AISBF proxy-awareness
+        proxy_set_header Host $host;
+        proxy_set_header X-Real-IP $remote_addr;
+        proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;
+        proxy_set_header X-Forwarded-Proto $scheme;
+        proxy_set_header X-Forwarded-Host $host;
+        proxy_set_header X-Forwarded-Port $server_port;
+        
+        # WebSocket support for streaming
+        proxy_set_header Upgrade $http_upgrade;
+        proxy_set_header Connection "upgrade";
+        
+        # Timeouts for long-running AI requests
+        proxy_connect_timeout 60s;
+        proxy_send_timeout 600s;
+        proxy_read_timeout 600s;
+        
+        # Buffering settings for streaming
+        proxy_buffering off;
+        proxy_cache off;
+        
+        # CORS headers (if needed)
+        add_header Access-Control-Allow-Origin "*" always;
+        add_header Access-Control-Allow-Methods "GET, POST, OPTIONS" always;
+        add_header Access-Control-Allow-Headers "Authorization, Content-Type" always;
+    }
+    
+    # Dashboard with stricter rate limiting
+    location /dashboard/ {
+        limit_req zone=aisbf_dashboard burst=10 nodelay;
+        
+        proxy_pass http://127.0.0.1:17765;
+        proxy_http_version 1.1;
+        
+        # Proxy headers for AISBF proxy-awareness
+        proxy_set_header Host $host;
+        proxy_set_header X-Real-IP $remote_addr;
+        proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;
+        proxy_set_header X-Forwarded-Proto $scheme;
+        proxy_set_header X-Forwarded-Host $host;
+        proxy_set_header X-Forwarded-Port $server_port;
+        
+        # Standard proxy settings
+        proxy_buffering off;
+    }
+    
+    # Block all other paths
+    location / {
+        return 404;
+    }
+    
+    # Security headers
+    add_header X-Frame-Options "SAMEORIGIN" always;
+    add_header X-Content-Type-Options "nosniff" always;
+    add_header X-XSS-Protection "1; mode=block" always;
+    
+    # Logging
+    access_log /var/log/nginx/aisbf_api_access.log;
+    error_log /var/log/nginx/aisbf_api_error.log;
+}
+```
+
+### Testing Proxy Configuration
+
+After configuring nginx, test the setup:
+
+1. **Test nginx configuration:**
+   ```bash
+   sudo nginx -t
+   ```
+
+2. **Reload nginx:**
+   ```bash
+   sudo systemctl reload nginx
+   ```
+
+3. **Test AISBF through proxy:**
+   ```bash
+   # Test root endpoint
+   curl https://aisbf.example.com/
+   
+   # Test with subpath
+   curl https://example.com/aisbf/
+   
+   # Test API endpoint
+   curl -X POST https://aisbf.example.com/api/v1/chat/completions \
+     -H "Content-Type: application/json" \
+     -d '{"model": "provider/model-name", "messages": [{"role": "user", "content": "Hello"}]}'
+   ```
+
+4. **Check dashboard URLs:**
+   - Open the dashboard in a browser
+   - Verify all links and redirects work correctly
+   - Check that the restart button uses the correct URL
+
+### Troubleshooting Proxy Issues
+
+If you encounter issues:
+
+1. **Check nginx error logs:**
+   ```bash
+   sudo tail -f /var/log/nginx/aisbf_error.log
+   ```
+
+2. **Verify proxy headers are being sent:**
+   - Check AISBF logs for received headers
+   - Enable debug logging in AISBF if needed
+
+3. **Common issues:**
+   - **404 errors with subpath**: Ensure `X-Forwarded-Prefix` header is set correctly
+   - **Redirect loops**: Check that `X-Forwarded-Proto` is set to `https`
+   - **Broken dashboard links**: Verify all proxy headers are configured
+   - **Streaming not working**: Ensure `proxy_buffering off` is set
+
+### Security Considerations
+
+When deploying behind a proxy:
+
+1. **Bind AISBF to localhost only** (default: `127.0.0.1:17765`)
+2. **Use HTTPS** for all external connections
+3. **Enable authentication** in AISBF configuration
+4. **Configure rate limiting** in nginx
+5. **Set security headers** in nginx configuration
+6. **Keep SSL certificates up to date**
+7. **Monitor access logs** for suspicious activity
+
 ## Key Classes and Functions
 
 ### aisbf/config.py

README.md

@@ -14,7 +14,10 @@ A modular proxy server for managing multiple AI provider integrations with unifi
 - **Token Rate Limiting**: Per-model token usage tracking with TPM (tokens per minute), TPH (tokens per hour), and TPD (tokens per day) limits
 - **Automatic Provider Disabling**: Providers automatically disabled when token rate limits are exceeded
 - **Context Management**: Automatic context condensation when approaching model limits with multiple condensation methods
+- **Provider-Level Defaults**: Set default condensation settings at provider level with cascading fallback logic
 - **Effective Context Tracking**: Reports total tokens used (effective_context) for every request
+- **SSL/TLS Support**: Built-in HTTPS support with Let's Encrypt integration and automatic certificate renewal
+- **Self-Signed Certificates**: Automatic generation of self-signed certificates for development/testing
 
 ## Author
 
@@ -76,6 +79,64 @@ See [`PYPI.md`](PYPI.md) for detailed instructions on publishing to PyPI.
 - Ollama (direct HTTP)
 ## Configuration
 
+### SSL/TLS Configuration
+
+AISBF supports HTTPS with automatic certificate management:
+
+#### Self-Signed Certificates (Default)
+- Automatically generated on first run when HTTPS is enabled
+- Stored in `~/.aisbf/cert.pem` and `~/.aisbf/key.pem`
+- Valid for 365 days
+- Suitable for development and internal use
+
+#### Let's Encrypt Integration
+Configure a public domain in the dashboard settings to enable Let's Encrypt:
+- Automatic certificate generation using certbot
+- Automatic renewal when certificates expire within 30 days
+- Valid certificates trusted by all browsers
+- Requires certbot to be installed on the system
+
+**Installation:**
+```bash
+# Ubuntu/Debian
+sudo apt-get install certbot
+
+# CentOS/RHEL
+sudo yum install certbot
+
+# macOS
+brew install certbot
+```
+
+**Configuration:**
+1. Navigate to Dashboard → Settings
+2. Set Protocol to "HTTPS"
+3. Enter your public domain (e.g., `api.example.com`)
+4. Optionally specify custom certificate/key paths
+5. Save settings
+
+The system will automatically:
+- Generate certificates using Let's Encrypt if a public domain is configured
+- Fall back to self-signed certificates if Let's Encrypt fails
+- Check certificate expiry on startup
+- Renew certificates when they expire within 30 days
+
+### Provider-Level Defaults
+
+Providers can now define default settings that cascade to all models:
+
+- **`default_context_size`**: Default maximum context size for all models in this provider
+- **`default_condense_context`**: Default condensation threshold percentage (0-100)
+- **`default_condense_method`**: Default condensation method(s) for all models
+
+**Cascading Priority:**
+1. Rotation model config (highest priority, only if explicitly set)
+2. Provider model-specific config
+3. Provider default config (NEW)
+4. System defaults
+
+This allows minimal model definitions in rotations - unspecified fields automatically inherit from provider defaults.
+
 ### Model Configuration
 
 Models can be configured with the following optional fields:
@@ -86,7 +147,7 @@ Models can be configured with the following optional fields:
 - **`rate_limit_TPD`**: Maximum tokens allowed per day (Tokens Per Day)
 - **`context_size`**: Maximum context size in tokens for the model. Used to determine when to trigger context condensation.
 - **`condense_context`**: Percentage (0-100) at which to trigger context condensation. 0 means disabled, any other value triggers condensation when context reaches this percentage of context_size.
-- **`condense_method`**: String or list of strings specifying condensation method(s). Supported values: "hierarchical", "conversational", "semantic", "algoritmic". Multiple methods can be chained together.
+- **`condense_method`**: String or list of strings specifying condensation method(s). Supported values: "hierarchical", "conversational", "semantic", "algorithmic". Multiple methods can be chained together.
 
 When token rate limits are exceeded, providers are automatically disabled:
 - TPM limit exceeded: Provider disabled for 1 minute
@@ -97,68 +158,133 @@ When token rate limits are exceeded, providers are automatically disabled:
 
 When context exceeds the configured percentage of `context_size`, the system automatically condenses the prompt using one or more methods:
 
-1. **Hierarchical**: Separates context into persistent (long-term facts) and transient (immediate task) layers
-2. **Conversational**: Summarizes old messages using a smaller model to maintain conversation continuity
-3. **Semantic**: Prunes irrelevant context based on current query using a smaller "janitor" model
-4. **Algoritmic**: Uses mathematical compression for technical data and logs (similar to LLMLingua)
+1. **Hierarchical**: Separates context into persistent (long-term facts) and transient (immediate task) layers - Pure algorithmic, no LLM needed
+2. **Conversational**: Summarizes old messages using a smaller model to maintain conversation continuity - LLM-based
+3. **Semantic**: Prunes irrelevant context based on current query using a smaller "janitor" model - LLM-based
+4. **Algorithmic**: Uses mathematical compression for technical data and logs (similar to LLMLingua) - Pure algorithmic, no LLM needed
+
+**Note:** Only `conversational` and `semantic` methods require LLM calls and use prompt files from `config/`. The `hierarchical` and `algorithmic` methods are pure algorithmic transformations.
 
 See `config/providers.json` and `config/rotations.json` for configuration examples.
 
 ## API Endpoints
 
+### Three Proxy Paths
+
+AISBF provides three ways to proxy AI models:
+
+#### PATH 1: Direct Provider Models
+Format: `{provider_id}/{model_name}`
+- Access specific models from configured providers
+- Example: `openai/gpt-4`, `gemini/gemini-2.0-flash`, `anthropic/claude-3-5-sonnet-20241022`, `kilotest/kilo/free`
+
+#### PATH 2: Rotations
+Format: `rotation/{rotation_name}`
+- Weighted load balancing across multiple providers
+- Automatic failover on errors
+- Example: `rotation/coding`, `rotation/general`
+
+#### PATH 3: Autoselect
+Format: `autoselect/{autoselect_name}`
+- AI-powered model selection based on content analysis
+- Automatic routing to specialized models
+- Example: `autoselect/autoselect`
+
 ### General Endpoints
 - `GET /` - Server status and provider list (includes providers, rotations, and autoselect)
+- `GET /api/models` - List all available models from all three proxy paths
+- `GET /api/v1/models` - OpenAI-compatible model listing (same as `/api/models`)
 
-### OpenAI-Compatible v1 Endpoints (Recommended)
-These endpoints follow the standard OpenAI API format with `provider/model` notation:
-
-- `POST /api/v1/chat/completions` - Chat completions (model format: `provider/model-name`)
-  - Example: `{"model": "openai/gpt-4", "messages": [...]}`
-  - Supports providers, rotations, and autoselect
-- `GET /api/v1/models` - List all available models from all providers
-- `POST /api/v1/audio/transcriptions` - Audio transcription (model format: `provider/model-name`)
-- `POST /api/v1/audio/speech` - Text-to-speech (model format: `provider/model-name`)
-- `POST /api/v1/images/generations` - Image generation (model format: `provider/model-name`)
-- `POST /api/v1/embeddings` - Text embeddings (model format: `provider/model-name`)
-- `GET /api/proxy/{content_id}` - Proxy generated content (images, audio, etc.)
+### Chat Completions
+
+#### Unified Endpoints (Recommended)
+These endpoints accept all three proxy path formats in the model field:
+
+**OpenAI-Compatible Format:**
+```bash
+POST /api/v1/chat/completions
+{
+  "model": "openai/gpt-4",           # PATH 1: Direct provider
+  "messages": [...]
+}
+
+POST /api/v1/chat/completions
+{
+  "model": "rotation/coding",        # PATH 2: Rotation
+  "messages": [...]
+}
+
+POST /api/v1/chat/completions
+{
+  "model": "autoselect/autoselect",  # PATH 3: Autoselect
+  "messages": [...]
+}
+```
+
+**Alternative Format:**
+```bash
+POST /api/{provider_id}/chat/completions
+# Where provider_id can be:
+# - A provider name (e.g., "openai", "gemini")
+# - A rotation name (e.g., "coding", "general")
+# - An autoselect name (e.g., "autoselect")
+```
 
-### Provider Endpoints (Legacy)
-- `POST /api/{provider_id}/chat/completions` - Chat completions for a specific provider
-- `GET /api/{provider_id}/models` - List available models for a specific provider
-- `POST /api/{provider_id}/audio/transcriptions` - Audio transcription
-- `POST /api/{provider_id}/audio/speech` - Text-to-speech
-- `POST /api/{provider_id}/images/generations` - Image generation
-- `POST /api/{provider_id}/embeddings` - Text embeddings
+### Audio Endpoints
+Model specified in request (supports all three proxy paths):
+
+- `POST /api/audio/transcriptions` - Audio transcription
+  - Example: `{"model": "openai/whisper-1", "file": ...}` or `{"model": "rotation/coding", "file": ...}`
+- `POST /api/audio/speech` - Text-to-speech
+  - Example: `{"model": "openai/tts-1", "input": "Hello"}` or `{"model": "rotation/coding", "input": "Hello"}`
+- `POST /api/v1/audio/transcriptions` - OpenAI-compatible audio transcription
+- `POST /api/v1/audio/speech` - OpenAI-compatible text-to-speech
+
+### Image Endpoints
+Model specified in request (supports all three proxy paths):
+
+- `POST /api/images/generations` - Image generation
+  - Example: `{"model": "openai/dall-e-3", "prompt": "A cat"}` or `{"model": "rotation/coding", "prompt": "A cat"}`
+- `POST /api/v1/images/generations` - OpenAI-compatible image generation
+
+### Embeddings Endpoints
+Model specified in request (supports all three proxy paths):
+
+- `POST /api/embeddings` - Text embeddings
+  - Example: `{"model": "openai/text-embedding-ada-002", "input": "Hello"}` or `{"model": "rotation/coding", "input": "Hello"}`
+- `POST /api/v1/embeddings` - OpenAI-compatible embeddings
+
+### Legacy Endpoints
+These endpoints are maintained for backward compatibility:
 
-### Rotation Endpoints
 - `GET /api/rotations` - List all available rotation configurations
-- `POST /api/rotations/chat/completions` - Chat completions using rotation (load balancing across providers)
-  - **Rotation Models**: Weighted random selection of models across multiple providers
-  - Automatic failover between providers on errors
-  - Configurable weights for each model to prioritize preferred options
-  - Supports both streaming and non-streaming responses
+- `POST /api/rotations/chat/completions` - Chat completions using rotation (model name = rotation name)
 - `GET /api/rotations/models` - List all models across all rotation configurations
-
-### Autoselect Endpoints
 - `GET /api/autoselect` - List all available autoselect configurations
-- `POST /api/autoselect/chat/completions` - Chat completions using AI-assisted selection based on content analysis
-  - **Autoselect Models**: AI analyzes request content to select the most appropriate model
-  - Automatic routing to specialized models based on task type (coding, analysis, creative writing, etc.)
-  - Fallback to default model if selection fails
-  - Supports both streaming and non-streaming responses
+- `POST /api/autoselect/chat/completions` - Chat completions using autoselect (model name = autoselect name)
 - `GET /api/autoselect/models` - List all models across all autoselect configurations
+- `GET /api/{provider_id}/models` - List available models for a specific provider, rotation, or autoselect
+
+### Content Proxy
+- `GET /api/proxy/{content_id}` - Proxy generated content (images, audio, etc.)
 
 ### Dashboard Endpoints
 - `GET /dashboard` - Web-based configuration dashboard
 - `GET /dashboard/login` - Dashboard login page
 - `POST /dashboard/login` - Handle dashboard authentication
 - `GET /dashboard/logout` - Logout from dashboard
-- `GET /dashboard/providers` - Edit providers configuration
+- `GET /dashboard/providers` - Edit providers configuration (includes provider-level defaults for condensation)
 - `GET /dashboard/rotations` - Edit rotations configuration
-- `GET /dashboard/autoselect` - Edit autoselect configuration
-- `GET /dashboard/settings` - Edit server settings
+- `GET /dashboard/autoselect` - Edit autoselect configuration (supports multiple autoselect configs with "internal" model option)
+- `GET /dashboard/settings` - Edit server settings (includes SSL/TLS configuration with Let's Encrypt support)
 - `POST /dashboard/restart` - Restart the server
 
+**Dashboard Features:**
+- **Provider Configuration**: Set default condensation settings at provider level that cascade to all models
+- **Autoselect Configuration**: Configure multiple autoselect models with "internal" option for using local HuggingFace models
+- **SSL/TLS Management**: Configure HTTPS with automatic Let's Encrypt certificate generation and renewal
+- **Collapsible UI**: All configuration sections use collapsible panels for better organization
+
 ## Error Handling
 - Rate limiting for failed requests
 - Automatic retry with provider rotation

aisbf.sh

@@ -91,15 +91,22 @@ ensure_venv() {
     fi
 }
 
-# Function to update venv packages silently
+# Function to update venv packages (only install missing ones, no forced upgrades)
 update_venv() {
-    # Update requirements if requirements.txt exists
+    # Only update if requirements file exists
     if [ -f "$SHARE_DIR/requirements.txt" ]; then
-        "$VENV_DIR/bin/pip" install --upgrade -r "$SHARE_DIR/requirements.txt" -q 2>/dev/null
-    fi
+        # Check if there are any new packages to install (not already satisfied)
+        "$VENV_DIR/bin/pip" install -r "$SHARE_DIR/requirements.txt" 2>&1 | grep -q "Requirement already satisfied"
+        ALREADY_SATISFIED=$?
         
-    # Update aisbf package silently
-    "$VENV_DIR/bin/pip" install --upgrade aisbf -q 2>/dev/null
+        if [ $ALREADY_SATISFIED -ne 0 ]; then
+            echo "Installing new requirements (this will take a while!) ... "
+            "$VENV_DIR/bin/pip" install -r "$SHARE_DIR/requirements.txt"
+            echo "[OK]"
+        else
+            echo "Virtual env already up to date"
+        fi
+    fi
 }
 
 # Function to start the server
@@ -121,14 +128,19 @@ start_server() {
     
     echo "Starting AISBF on port $PORT..."
     
-    # Start the proxy server with logging
-    # Redirect stderr to suppress BrokenPipeError during shutdown
-    uvicorn main:app --host 127.0.0.1 --port $PORT 2>&1 | while IFS= read -r line; do
-        # Filter out BrokenPipeError logging errors
-        if [[ "$line" != *"--- Logging error ---"* ]] && [[ "$line" != *"BrokenPipeError"* ]] && [[ "$line" != *"Call stack:"* ]] && [[ "$line" != *"File "*"/python"* ]] && [[ "$line" != *"Message:"* ]] && [[ "$line" != *"Arguments:"* ]]; then
-            echo "$line" | tee -a "$LOG_DIR/aisbf_stdout.log"
+    # Check if debug mode is enabled
+    if [ "$DEBUG" = "true" ]; then
+        echo "Debug mode enabled - showing all debug messages"
+        export AISBF_DEBUG=true
+    fi
+    
+    # Start the proxy server - runs in foreground
+    # Use exec to replace the shell process so signals are properly handled
+    if [ "$DEBUG" = "true" ]; then
+        exec uvicorn main:app --host 127.0.0.1 --port $PORT --log-level debug 2>&1 | tee -a "$LOG_DIR/aisbf_stdout.log"
+    else
+        exec uvicorn main:app --host 127.0.0.1 --port $PORT 2>&1 | grep -v -E "(--- Logging error ---|BrokenPipeError|Call stack:|Message:|Arguments:)" | tee -a "$LOG_DIR/aisbf_stdout.log"
     fi
-    done
 }
 
 # Function to start as daemon
@@ -156,9 +168,19 @@ start_daemon() {
     
     echo "Starting AISBF on port $PORT in background..."
     
+    # Check if debug mode is enabled
+    if [ "$DEBUG" = "true" ]; then
+        echo "Debug mode enabled - showing all debug messages"
+        export AISBF_DEBUG=true
+    fi
+    
     # Start in background with nohup and logging
     # Filter out BrokenPipeError logging errors
+    if [ "$DEBUG" = "true" ]; then
+        nohup bash -c "source $VENV_DIR/bin/activate && cd $SHARE_DIR && uvicorn main:app --host 127.0.0.1 --port $PORT --log-level debug 2>&1" >> "$LOG_DIR/aisbf_stdout.log" 2>&1 &
+    else
         nohup bash -c "source $VENV_DIR/bin/activate && cd $SHARE_DIR && uvicorn main:app --host 127.0.0.1 --port $PORT 2>&1 | grep -v '--- Logging error ---' | grep -v 'BrokenPipeError' | grep -v 'Call stack:' | grep -v 'File .*python' | grep -v 'Message:' | grep -v 'Arguments:'" >> "$LOG_DIR/aisbf_stdout.log" 2>&1 &
+    fi
     PID=$!
     echo $PID > "$PIDFILE"
     echo "AISBF started in background (PID: $PID)"
@@ -200,8 +222,59 @@ stop_daemon() {
     fi
 }
 
+# Function to show help
+show_help() {
+    echo "AISBF - AI Service Broker Framework"
+    echo ""
+    echo "Usage: aisbf.sh [OPTIONS] [COMMAND]"
+    echo ""
+    echo "Options:"
+    echo "  --debug     Enable debug mode with verbose logging"
+    echo "  -h, --help  Show this help message"
+    echo ""
+    echo "Commands:"
+    echo "  daemon   Start AISBF in background (daemon mode)"
+    echo "  status   Check if AISBF is running"
+    echo "  stop     Stop the AISBF daemon"
+    echo ""
+    echo "Examples:"
+    echo "  aisbf.sh                  # Start in foreground"
+    echo "  aisbf.sh --debug          # Start with debug logging"
+    echo "  aisbf.sh daemon           # Start in background"
+    echo "  aisbf.sh --debug daemon   # Start in background with debug"
+    echo "  aisbf.sh status           # Check status"
+    echo "  aisbf.sh stop             # Stop the server"
+}
+
+# Parse command line arguments
+DEBUG="false"
+COMMAND=""
+
+while [[ $# -gt 0 ]]; do
+    case $1 in
+        --debug|-d)
+            DEBUG="true"
+            shift
+            ;;
+        -h|--help)
+            show_help
+            exit 0
+            ;;
+        daemon|status|stop)
+            COMMAND="$1"
+            shift
+            ;;
+        *)
+            echo "Unknown option: $1"
+            echo ""
+            show_help
+            exit 1
+            ;;
+    esac
+done
+
 # Main command handling
-case "$1" in
+case "$COMMAND" in
     daemon)
         start_daemon
         ;;

aisbf/config.py

@@ -66,6 +66,7 @@ class ProviderConfig(BaseModel):
     default_condense_method: Optional[Union[str, List[str]]] = None
 
 class RotationConfig(BaseModel):
+    model_name: str
     providers: List[Dict]
     notifyerrors: bool = False
     # Default settings for models in this rotation
@@ -104,12 +105,16 @@ class Config:
         if custom_dir:
             self._custom_config_dir = Path(custom_dir)
         
+        # Track loaded file paths for summary
+        self._loaded_files = {}
+        
         self._ensure_config_directory()
         self._load_providers()
         self._load_rotations()
         self._load_autoselect()
         self._load_condensation()
         self._initialize_error_tracking()
+        self._log_configuration_summary()
 
     def _get_config_source_dir(self):
         """Get the directory containing default config files"""
@@ -187,6 +192,7 @@ class Config:
         with open(providers_path) as f:
             data = json.load(f)
             self.providers = {k: ProviderConfig(**v) for k, v in data['providers'].items()}
+            self._loaded_files['providers'] = str(providers_path.absolute())
             logger.info(f"Loaded {len(self.providers)} providers: {list(self.providers.keys())}")
             for provider_id, provider_config in self.providers.items():
                 logger.info(f"  - {provider_id}: type={provider_config.type}, endpoint={provider_config.endpoint}")
@@ -214,6 +220,7 @@ class Config:
         logger.info(f"Loading rotations from: {rotations_path}")
         with open(rotations_path) as f:
             data = json.load(f)
+            self._loaded_files['rotations'] = str(rotations_path.absolute())
             
             # Extract global notifyerrors setting (top-level, outside rotations)
             self.global_notifyerrors = data.get('notifyerrors', False)
@@ -254,18 +261,31 @@ class Config:
             logger.info(f"=== Config._load_rotations END ===")
 
     def _load_autoselect(self):
+        import logging
+        logger = logging.getLogger(__name__)
+        logger.info(f"=== Config._load_autoselect START ===")
+        
         autoselect_path = Path.home() / '.aisbf' / 'autoselect.json'
+        logger.info(f"Looking for autoselect at: {autoselect_path}")
+        
         if not autoselect_path.exists():
+            logger.info(f"User config not found, falling back to source config")
             # Fallback to source config if user config doesn't exist
             try:
                 source_dir = self._get_config_source_dir()
                 autoselect_path = source_dir / 'autoselect.json'
+                logger.info(f"Using source config at: {autoselect_path}")
             except FileNotFoundError:
+                logger.error("Could not find autoselect.json configuration file")
                 raise FileNotFoundError("Could not find autoselect.json configuration file")
         
+        logger.info(f"Loading autoselect from: {autoselect_path}")
         with open(autoselect_path) as f:
             data = json.load(f)
             self.autoselect = {k: AutoselectConfig(**v) for k, v in data.items()}
+            self._loaded_files['autoselect'] = str(autoselect_path.absolute())
+            logger.info(f"Loaded {len(self.autoselect)} autoselect configurations: {list(self.autoselect.keys())}")
+            logger.info(f"=== Config._load_autoselect END ===")
     
     def _load_condensation(self):
         """Load condensation configuration from providers.json"""
@@ -274,20 +294,26 @@ class Config:
         logger.info(f"=== Config._load_condensation START ===")
         
         providers_path = Path.home() / '.aisbf' / 'providers.json'
+        logger.info(f"Looking for condensation config in: {providers_path}")
+        
         if not providers_path.exists():
+            logger.info(f"User config not found, falling back to source config")
             # Fallback to source config if user config doesn't exist
             try:
                 source_dir = self._get_config_source_dir()
                 providers_path = source_dir / 'providers.json'
+                logger.info(f"Using source config at: {providers_path}")
             except FileNotFoundError:
                 logger.warning("Could not find providers.json for condensation config")
                 self.condensation = CondensationConfig()
                 return
         
+        logger.info(f"Loading condensation config from: {providers_path}")
         with open(providers_path) as f:
             data = json.load(f)
             condensation_data = data.get('condensation', {})
             self.condensation = CondensationConfig(**condensation_data)
+            self._loaded_files['condensation'] = str(providers_path.absolute())
             logger.info(f"Loaded condensation config: provider_id={self.condensation.provider_id}, model={self.condensation.model}, enabled={self.condensation.enabled}")
             logger.info(f"=== Config._load_condensation END ===")
 
@@ -300,6 +326,31 @@ class Config:
                 'disabled_until': None
             }
     
+    def _log_configuration_summary(self):
+        """Log a summary of all loaded configuration files"""
+        import logging
+        logger = logging.getLogger(__name__)
+        
+        logger.info("")
+        logger.info("=" * 80)
+        logger.info("=== CONFIGURATION FILES LOADED ===")
+        logger.info("=" * 80)
+        
+        if 'providers' in self._loaded_files:
+            logger.info(f"Providers:    {self._loaded_files['providers']}")
+        
+        if 'rotations' in self._loaded_files:
+            logger.info(f"Rotations:    {self._loaded_files['rotations']}")
+        
+        if 'autoselect' in self._loaded_files:
+            logger.info(f"Autoselect:   {self._loaded_files['autoselect']}")
+        
+        if 'condensation' in self._loaded_files:
+            logger.info(f"Condensation: {self._loaded_files['condensation']}")
+        
+        logger.info("=" * 80)
+        logger.info("")
+
     def get_provider(self, provider_id: str) -> ProviderConfig:
         import logging
         logger = logging.getLogger(__name__)

aisbf/context.py

@@ -145,6 +145,8 @@ class ContextManager:
     def _initialize_internal_model(self):
         """Initialize the internal HuggingFace model for condensation (lazy loading)"""
         import logging
+        import json
+        from pathlib import Path
         logger = logging.getLogger(__name__)
         
         if self._internal_model is not None:
@@ -156,7 +158,33 @@ class ContextManager:
             import threading
             
             logger.info("=== INITIALIZING INTERNAL CONDENSATION MODEL ===")
-            model_name = "huihui-ai/Qwen2.5-0.5B-Instruct-abliterated-v3"
+            
+            # Load model name from config
+            config_path = Path.home() / '.aisbf' / 'aisbf.json'
+            if not config_path.exists():
+                # Try installed locations
+                installed_dirs = [
+                    Path('/usr/share/aisbf'),
+                    Path.home() / '.local' / 'share' / 'aisbf',
+                ]
+                for installed_dir in installed_dirs:
+                    test_path = installed_dir / 'aisbf.json'
+                    if test_path.exists():
+                        config_path = test_path
+                        break
+                else:
+                    # Fallback to source tree
+                    config_path = Path(__file__).parent.parent / 'config' / 'aisbf.json'
+            
+            model_name = "huihui-ai/Qwen2.5-0.5B-Instruct-abliterated-v3"  # Default
+            if config_path.exists():
+                try:
+                    with open(config_path) as f:
+                        aisbf_config = json.load(f)
+                        model_name = aisbf_config.get('internal_model', {}).get('condensation_model_id', model_name)
+                except Exception as e:
+                    logger.warning(f"Error loading condensation model config: {e}, using default")
+            
             logger.info(f"Model: {model_name}")
             
             # Check for GPU availability
@@ -700,7 +728,13 @@ def get_context_config_for_model(
     rotation_model_config: Optional[Dict] = None
 ) -> Dict:
     """
-    Get context configuration for a specific model.
+    Get context configuration for a specific model with cascading fallback.
+    
+    Priority order for each field:
+    1. Rotation model config (if explicitly set)
+    2. Provider model-specific config (if exists)
+    3. Provider default config (fallback)
+    4. System default (0 for condense_context, None for others)
     
     Args:
         model_name: Name of the model
@@ -716,19 +750,47 @@ def get_context_config_for_model(
         'condense_method': None
     }
     
-    # Check rotation model config first (highest priority)
-    if rotation_model_config:
-        context_config['context_size'] = rotation_model_config.get('context_size')
-        context_config['condense_context'] = rotation_model_config.get('condense_context', 0)
-        context_config['condense_method'] = rotation_model_config.get('condense_method')
-    
-    # Fall back to provider config
-    elif provider_config and hasattr(provider_config, 'models'):
+    # Step 1: Get provider-level defaults and model-specific config
+    model_specific_config = None
+    if provider_config:
+        # Try to find model-specific config in provider
+        if hasattr(provider_config, 'models') and provider_config.models:
             for model in provider_config.models:
                 if model.get('name') == model_name:
-                context_config['context_size'] = model.get('context_size')
-                context_config['condense_context'] = model.get('condense_context', 0)
-                context_config['condense_method'] = model.get('condense_method')
+                    model_specific_config = model
                     break
         
+        # Build base config from provider (model-specific > provider defaults)
+        # context_size
+        if model_specific_config and model_specific_config.get('context_size') is not None:
+            context_config['context_size'] = model_specific_config.get('context_size')
+        elif hasattr(provider_config, 'default_context_size') and provider_config.default_context_size is not None:
+            context_config['context_size'] = provider_config.default_context_size
+        
+        # condense_context
+        if model_specific_config and model_specific_config.get('condense_context') is not None:
+            context_config['condense_context'] = model_specific_config.get('condense_context')
+        elif hasattr(provider_config, 'default_condense_context') and provider_config.default_condense_context is not None:
+            context_config['condense_context'] = provider_config.default_condense_context
+        else:
+            context_config['condense_context'] = 0
+        
+        # condense_method
+        if model_specific_config and model_specific_config.get('condense_method') is not None:
+            context_config['condense_method'] = model_specific_config.get('condense_method')
+        elif hasattr(provider_config, 'default_condense_method') and provider_config.default_condense_method is not None:
+            context_config['condense_method'] = provider_config.default_condense_method
+    
+    # Step 2: Override with rotation model config (only if explicitly set)
+    if rotation_model_config:
+        # Only override if the field is explicitly set in rotation config
+        if 'context_size' in rotation_model_config and rotation_model_config['context_size'] is not None:
+            context_config['context_size'] = rotation_model_config['context_size']
+        
+        if 'condense_context' in rotation_model_config and rotation_model_config['condense_context'] is not None:
+            context_config['condense_context'] = rotation_model_config['condense_context']
+        
+        if 'condense_method' in rotation_model_config and rotation_model_config['condense_method'] is not None:
+            context_config['condense_method'] = rotation_model_config['condense_method']
+    
     return context_config
\ No newline at end of file

aisbf/handlers.py

@@ -1366,6 +1366,11 @@ class RotationHandler:
             
             logger.info(f"  [AVAILABLE] Provider {provider_id} is active and ready")
             
+            # Check if provider-level weight is specified
+            provider_weight = provider.get('weight', 1)  # Default to 1 if not specified
+            if provider.get('weight') is not None:
+                logger.info(f"  Provider-level weight: {provider_weight}")
+            
             # Check if models are specified in rotation config
             # If not, use models from provider config
             rotation_models = provider.get('models')
@@ -1375,17 +1380,17 @@ class RotationHandler:
                 
                 # Get models from provider config
                 if provider_config.models:
-                    # Use models from provider config with default weight of 1
+                    # Use models from provider config with provider-level weight
                     rotation_models = []
                     for provider_model in provider_config.models:
                         model_dict = {
                             'name': provider_model.name,
-                            'weight': 1,  # Default weight
+                            'weight': provider_weight,  # Use provider-level weight
                             'rate_limit': provider_model.rate_limit,
                             'max_request_tokens': provider_model.max_request_tokens
                         }
                         rotation_models.append(model_dict)
-                    logger.info(f"  Loaded {len(rotation_models)} model(s) from provider config")
+                    logger.info(f"  Loaded {len(rotation_models)} model(s) from provider config with weight {provider_weight}")
                 else:
                     logger.warning(f"  No models defined in provider config for {provider_id}")
                     logger.warning(f"  Skipping this provider")
@@ -2530,6 +2535,8 @@ class AutoselectHandler:
     def _initialize_internal_model(self):
         """Initialize the internal HuggingFace model for selection (lazy loading)"""
         import logging
+        import json
+        from pathlib import Path
         logger = logging.getLogger(__name__)
         
         if self._internal_model is not None:
@@ -2541,7 +2548,33 @@ class AutoselectHandler:
             import threading
             
             logger.info("=== INITIALIZING INTERNAL SELECTION MODEL ===")
-            model_name = "huihui-ai/Qwen2.5-0.5B-Instruct-abliterated-v3"
+            
+            # Load model name from config
+            config_path = Path.home() / '.aisbf' / 'aisbf.json'
+            if not config_path.exists():
+                # Try installed locations
+                installed_dirs = [
+                    Path('/usr/share/aisbf'),
+                    Path.home() / '.local' / 'share' / 'aisbf',
+                ]
+                for installed_dir in installed_dirs:
+                    test_path = installed_dir / 'aisbf.json'
+                    if test_path.exists():
+                        config_path = test_path
+                        break
+                else:
+                    # Fallback to source tree
+                    config_path = Path(__file__).parent.parent / 'config' / 'aisbf.json'
+            
+            model_name = "huihui-ai/Qwen2.5-0.5B-Instruct-abliterated-v3"  # Default
+            if config_path.exists():
+                try:
+                    with open(config_path) as f:
+                        aisbf_config = json.load(f)
+                        model_name = aisbf_config.get('internal_model', {}).get('autoselect_model_id', model_name)
+                except Exception as e:
+                    logger.warning(f"Error loading autoselect model config: {e}, using default")
+            
             logger.info(f"Model: {model_name}")
             
             # Check for GPU availability

aisbf/utils.py

@@ -178,3 +178,334 @@ def get_max_request_tokens_for_model(
     
     logger.debug(f"No max_request_tokens configured for model {model_name}")
     return None
+
+
+def generate_self_signed_certificate(cert_path: str, key_path: str) -> bool:
+    """
+    Generate a self-signed SSL certificate and private key.
+    
+    Args:
+        cert_path: Path where the certificate should be saved
+        key_path: Path where the private key should be saved
+    
+    Returns:
+        True if successful, False otherwise
+    """
+    import logging
+    from pathlib import Path
+    logger = logging.getLogger(__name__)
+    
+    try:
+        from cryptography import x509
+        from cryptography.x509.oid import NameOID
+        from cryptography.hazmat.primitives import hashes
+        from cryptography.hazmat.primitives.asymmetric import rsa
+        from cryptography.hazmat.primitives import serialization
+        import datetime
+        
+        logger.info(f"Generating self-signed SSL certificate...")
+        logger.info(f"Certificate path: {cert_path}")
+        logger.info(f"Key path: {key_path}")
+        
+        # Generate private key
+        private_key = rsa.generate_private_key(
+            public_exponent=65537,
+            key_size=2048,
+        )
+        
+        # Create certificate
+        subject = issuer = x509.Name([
+            x509.NameAttribute(NameOID.COUNTRY_NAME, "US"),
+            x509.NameAttribute(NameOID.STATE_OR_PROVINCE_NAME, "State"),
+            x509.NameAttribute(NameOID.LOCALITY_NAME, "City"),
+            x509.NameAttribute(NameOID.ORGANIZATION_NAME, "AISBF"),
+            x509.NameAttribute(NameOID.COMMON_NAME, "localhost"),
+        ])
+        
+        cert = x509.CertificateBuilder().subject_name(
+            subject
+        ).issuer_name(
+            issuer
+        ).public_key(
+            private_key.public_key()
+        ).serial_number(
+            x509.random_serial_number()
+        ).not_valid_before(
+            datetime.datetime.utcnow()
+        ).not_valid_after(
+            datetime.datetime.utcnow() + datetime.timedelta(days=365)
+        ).add_extension(
+            x509.SubjectAlternativeName([
+                x509.DNSName("localhost"),
+                x509.DNSName("127.0.0.1"),
+            ]),
+            critical=False,
+        ).sign(private_key, hashes.SHA256())
+        
+        # Ensure directories exist
+        Path(cert_path).parent.mkdir(parents=True, exist_ok=True)
+        Path(key_path).parent.mkdir(parents=True, exist_ok=True)
+        
+        # Write certificate to file
+        with open(cert_path, "wb") as f:
+            f.write(cert.public_bytes(serialization.Encoding.PEM))
+        
+        # Write private key to file
+        with open(key_path, "wb") as f:
+            f.write(private_key.private_bytes(
+                encoding=serialization.Encoding.PEM,
+                format=serialization.PrivateFormat.TraditionalOpenSSL,
+                encryption_algorithm=serialization.NoEncryption()
+            ))
+        
+        logger.info(f"Successfully generated self-signed SSL certificate")
+        logger.info(f"Certificate valid for 365 days")
+        return True
+        
+    except ImportError:
+        logger.error("cryptography library not installed. Cannot generate self-signed certificate.")
+        logger.error("Install with: pip install cryptography")
+        return False
+    except Exception as e:
+        logger.error(f"Error generating self-signed certificate: {e}", exc_info=True)
+        return False
+
+
+def check_certificate_expiry(cert_path: str) -> Optional[int]:
+    """
+    Check how many days until certificate expires.
+    
+    Args:
+        cert_path: Path to certificate file
+    
+    Returns:
+        Number of days until expiry, or None if cannot be determined
+    """
+    import logging
+    from pathlib import Path
+    logger = logging.getLogger(__name__)
+    
+    try:
+        from cryptography import x509
+        from cryptography.hazmat.backends import default_backend
+        import datetime
+        
+        if not Path(cert_path).exists():
+            return None
+        
+        with open(cert_path, 'rb') as f:
+            cert_data = f.read()
+        
+        cert = x509.load_pem_x509_certificate(cert_data, default_backend())
+        expiry_date = cert.not_valid_after
+        days_until_expiry = (expiry_date - datetime.datetime.utcnow()).days
+        
+        logger.info(f"Certificate expires in {days_until_expiry} days")
+        return days_until_expiry
+        
+    except Exception as e:
+        logger.error(f"Error checking certificate expiry: {e}")
+        return None
+
+
+def generate_letsencrypt_certificate(domain: str, cert_path: str, key_path: str, email: Optional[str] = None) -> bool:
+    """
+    Generate Let's Encrypt SSL certificate using certbot.
+    
+    Args:
+        domain: Public domain name
+        cert_path: Path where certificate should be saved
+        key_path: Path where private key should be saved
+        email: Optional email for Let's Encrypt notifications
+    
+    Returns:
+        True if successful, False otherwise
+    """
+    import logging
+    import subprocess
+    from pathlib import Path
+    logger = logging.getLogger(__name__)
+    
+    try:
+        logger.info(f"Generating Let's Encrypt certificate for domain: {domain}")
+        
+        # Check if certbot is installed
+        try:
+            subprocess.run(['certbot', '--version'], capture_output=True, check=True)
+        except (subprocess.CalledProcessError, FileNotFoundError):
+            logger.error("certbot not found. Please install certbot:")
+            logger.error("  Ubuntu/Debian: sudo apt-get install certbot")
+            logger.error("  CentOS/RHEL: sudo yum install certbot")
+            logger.error("  macOS: brew install certbot")
+            return False
+        
+        # Prepare certbot command
+        config_dir = Path.home() / '.aisbf' / 'letsencrypt'
+        config_dir.mkdir(parents=True, exist_ok=True)
+        
+        cmd = [
+            'certbot', 'certonly',
+            '--standalone',
+            '--non-interactive',
+            '--agree-tos',
+            '--domain', domain,
+            '--config-dir', str(config_dir),
+            '--work-dir', str(config_dir / 'work'),
+            '--logs-dir', str(config_dir / 'logs'),
+        ]
+        
+        if email:
+            cmd.extend(['--email', email])
+        else:
+            cmd.append('--register-unsafely-without-email')
+        
+        logger.info(f"Running certbot command...")
+        result = subprocess.run(cmd, capture_output=True, text=True)
+        
+        if result.returncode != 0:
+            logger.error(f"certbot failed: {result.stderr}")
+            return False
+        
+        # Copy certificates to specified paths
+        le_cert = config_dir / 'live' / domain / 'fullchain.pem'
+        le_key = config_dir / 'live' / domain / 'privkey.pem'
+        
+        if le_cert.exists() and le_key.exists():
+            import shutil
+            shutil.copy2(le_cert, cert_path)
+            shutil.copy2(le_key, key_path)
+            logger.info(f"Let's Encrypt certificate generated successfully")
+            logger.info(f"Certificate valid for 90 days")
+            return True
+        else:
+            logger.error(f"Certificate files not found after certbot execution")
+            return False
+            
+    except Exception as e:
+        logger.error(f"Error generating Let's Encrypt certificate: {e}", exc_info=True)
+        return False
+
+
+def renew_letsencrypt_certificate(domain: str, cert_path: str, key_path: str) -> bool:
+    """
+    Renew Let's Encrypt SSL certificate.
+    
+    Args:
+        domain: Public domain name
+        cert_path: Path where certificate should be saved
+        key_path: Path where private key should be saved
+    
+    Returns:
+        True if successful, False otherwise
+    """
+    import logging
+    import subprocess
+    from pathlib import Path
+    logger = logging.getLogger(__name__)
+    
+    try:
+        logger.info(f"Renewing Let's Encrypt certificate for domain: {domain}")
+        
+        config_dir = Path.home() / '.aisbf' / 'letsencrypt'
+        
+        cmd = [
+            'certbot', 'renew',
+            '--config-dir', str(config_dir),
+            '--work-dir', str(config_dir / 'work'),
+            '--logs-dir', str(config_dir / 'logs'),
+            '--non-interactive',
+        ]
+        
+        result = subprocess.run(cmd, capture_output=True, text=True)
+        
+        if result.returncode != 0:
+            logger.error(f"certbot renew failed: {result.stderr}")
+            return False
+        
+        # Copy renewed certificates
+        le_cert = config_dir / 'live' / domain / 'fullchain.pem'
+        le_key = config_dir / 'live' / domain / 'privkey.pem'
+        
+        if le_cert.exists() and le_key.exists():
+            import shutil
+            shutil.copy2(le_cert, cert_path)
+            shutil.copy2(le_key, key_path)
+            logger.info(f"Let's Encrypt certificate renewed successfully")
+            return True
+        else:
+            logger.error(f"Renewed certificate files not found")
+            return False
+            
+    except Exception as e:
+        logger.error(f"Error renewing Let's Encrypt certificate: {e}", exc_info=True)
+        return False
+
+
+def ensure_ssl_certificates(cert_path: Optional[str] = None, key_path: Optional[str] = None,
+                           public_domain: Optional[str] = None, email: Optional[str] = None) -> tuple:
+    """
+    Ensure SSL certificates exist, generating them if necessary.
+    Uses Let's Encrypt if public_domain is provided, otherwise self-signed.
+    Also checks expiry and renews if needed.
+    
+    Args:
+        cert_path: Path to certificate file (None for default)
+        key_path: Path to key file (None for default)
+        public_domain: Public domain for Let's Encrypt (None for self-signed)
+        email: Email for Let's Encrypt notifications (optional)
+    
+    Returns:
+        Tuple of (cert_path, key_path) with resolved paths
+    """
+    import logging
+    from pathlib import Path
+    logger = logging.getLogger(__name__)
+    
+    # Use default paths if not specified
+    config_dir = Path.home() / '.aisbf'
+    if not cert_path:
+        cert_path = str(config_dir / 'cert.pem')
+    if not key_path:
+        key_path = str(config_dir / 'key.pem')
+    
+    # Expand user paths
+    cert_path = str(Path(cert_path).expanduser())
+    key_path = str(Path(key_path).expanduser())
+    
+    # Check if certificates exist
+    cert_exists = Path(cert_path).exists()
+    key_exists = Path(key_path).exists()
+    
+    # Check certificate expiry if it exists
+    needs_renewal = False
+    if cert_exists:
+        days_until_expiry = check_certificate_expiry(cert_path)
+        if days_until_expiry is not None and days_until_expiry < 30:
+            logger.warning(f"Certificate expires in {days_until_expiry} days, renewal needed")
+            needs_renewal = True
+    
+    # Generate or renew certificate
+    if not cert_exists or not key_exists or needs_renewal:
+        if public_domain:
+            # Try Let's Encrypt
+            logger.info(f"Using Let's Encrypt for domain: {public_domain}")
+            
+            if needs_renewal:
+                success = renew_letsencrypt_certificate(public_domain, cert_path, key_path)
+            else:
+                success = generate_letsencrypt_certificate(public_domain, cert_path, key_path, email)
+            
+            if not success:
+                logger.warning(f"Let's Encrypt failed, falling back to self-signed certificate")
+                if not generate_self_signed_certificate(cert_path, key_path):
+                    raise FileNotFoundError(f"Failed to generate SSL certificates")
+        else:
+            # Generate self-signed certificate
+            logger.info(f"Generating self-signed SSL certificate...")
+            if not generate_self_signed_certificate(cert_path, key_path):
+                raise FileNotFoundError(f"Failed to generate self-signed SSL certificate")
+    else:
+        logger.info(f"Using existing SSL certificate: {cert_path}")
+        logger.info(f"Using existing SSL key: {key_path}")
+    
+    return cert_path, key_path
\ No newline at end of file

config/aisbf.json

@@ -19,6 +19,7 @@
     "password": "8c6976e5b5410415bde908bd4dee15dfb167a9c873fc4bb8a81f6f2ab448a918"
   },
   "internal_model": {
-    "model_id": "huihui-ai/Qwen2.5-0.5B-Instruct-abliterated-v3"
+    "condensation_model_id": "huihui-ai/Qwen2.5-0.5B-Instruct-abliterated-v3",
+    "autoselect_model_id": "huihui-ai/Qwen2.5-0.5B-Instruct-abliterated-v3"
   }
 }

config/rotations.json

@@ -176,6 +176,27 @@
           ]
         }
       ]
+    },
+    "simple-gemini": {
+      "_comment": "SIMPLE FORMAT: Specify provider_id only, models will be loaded from providers.json automatically",
+      "model_name": "simple-gemini",
+      "notifyerrors": false,
+      "providers": [
+        {
+          "provider_id": "gemini"
+        }
+      ]
+    },
+    "simple-openai": {
+      "_comment": "SIMPLE FORMAT: Provider with custom weight, models loaded from providers.json",
+      "model_name": "simple-openai",
+      "notifyerrors": false,
+      "providers": [
+        {
+          "provider_id": "openai",
+          "weight": 2
+        }
+      ]
     }
   }
 }
\ No newline at end of file

main.py

@@ -31,6 +31,8 @@ from aisbf.models import ChatCompletionRequest, ChatCompletionResponse
 from aisbf.handlers import RequestHandler, RotationHandler, AutoselectHandler
 from aisbf.database import initialize_database
 from starlette.middleware.sessions import SessionMiddleware
+from starlette.middleware.base import BaseHTTPMiddleware
+from starlette.datastructures import Headers
 from itsdangerous import URLSafeTimedSerializer
 import time
 import logging
@@ -39,11 +41,14 @@ import os
 import argparse
 import secrets
 import hashlib
+import asyncio
 from logging.handlers import RotatingFileHandler
 from datetime import datetime, timedelta
 from collections import defaultdict
 from pathlib import Path
 import json
+import markdown
+from urllib.parse import urljoin
 
 # Global variable to store custom config directory
 _custom_config_dir = None
@@ -346,34 +351,413 @@ def setup_logging():
 # Configure logging
 logger = setup_logging()
 
+class ProxyHeadersMiddleware(BaseHTTPMiddleware):
+    """
+    Middleware to handle proxy headers and make the application proxy-aware.
+    
+    Supports standard proxy headers:
+    - X-Forwarded-Proto: Original protocol (http/https)
+    - X-Forwarded-Host: Original host
+    - X-Forwarded-Port: Original port
+    - X-Forwarded-Prefix or X-Script-Name: URL prefix/subpath
+    - X-Forwarded-For: Client IP address
+    """
+    
+    async def dispatch(self, request: Request, call_next):
+        # Get proxy headers
+        forwarded_proto = request.headers.get("X-Forwarded-Proto")
+        forwarded_host = request.headers.get("X-Forwarded-Host")
+        forwarded_port = request.headers.get("X-Forwarded-Port")
+        forwarded_prefix = request.headers.get("X-Forwarded-Prefix") or request.headers.get("X-Script-Name")
+        forwarded_for = request.headers.get("X-Forwarded-For")
+        
+        # Update request scope with proxy information
+        if forwarded_proto:
+            request.scope["scheme"] = forwarded_proto
+        
+        if forwarded_host:
+            # Handle host:port format
+            if ":" in forwarded_host and not forwarded_port:
+                host_parts = forwarded_host.split(":", 1)
+                request.scope["server"] = (host_parts[0], int(host_parts[1]))
+            else:
+                port = int(forwarded_port) if forwarded_port else (443 if forwarded_proto == "https" else 80)
+                request.scope["server"] = (forwarded_host, port)
+        elif forwarded_port:
+            # Only port was forwarded, keep existing host
+            current_host = request.scope.get("server", ("localhost", 80))[0]
+            request.scope["server"] = (current_host, int(forwarded_port))
+        
+        # Handle URL prefix/subpath
+        if forwarded_prefix:
+            # Remove trailing slash from prefix
+            forwarded_prefix = forwarded_prefix.rstrip("/")
+            request.scope["root_path"] = forwarded_prefix
+            
+            # Update path to remove prefix if present
+            original_path = request.scope.get("path", "")
+            if original_path.startswith(forwarded_prefix):
+                request.scope["path"] = original_path[len(forwarded_prefix):] or "/"
+        
+        # Store client IP from X-Forwarded-For
+        if forwarded_for:
+            # X-Forwarded-For can contain multiple IPs, take the first one (original client)
+            client_ip = forwarded_for.split(",")[0].strip()
+            request.scope["client"] = (client_ip, request.scope.get("client", ("", 0))[1])
+        
+        response = await call_next(request)
+        return response
+
+def get_base_url(request: Request) -> str:
+    """
+    Get the base URL for the application, respecting proxy headers.
+    
+    Returns the full base URL including scheme, host, port, and prefix.
+    Example: https://example.com:8443/aisbf
+    """
+    scheme = request.scope.get("scheme", "http")
+    server = request.scope.get("server", ("localhost", 80))
+    host = server[0]
+    port = server[1]
+    root_path = request.scope.get("root_path", "")
+    
+    # Don't include port in URL if it's the default for the scheme
+    if (scheme == "http" and port == 80) or (scheme == "https" and port == 443):
+        base_url = f"{scheme}://{host}{root_path}"
+    else:
+        base_url = f"{scheme}://{host}:{port}{root_path}"
+    
+    return base_url
+
+def url_for(request: Request, path: str) -> str:
+    """
+    Generate a proxy-aware URL for the given path.
+    
+    Args:
+        request: The current request object
+        path: The path to generate URL for (should start with /)
+    
+    Returns:
+        Full URL respecting proxy configuration
+    """
+    base_url = get_base_url(request)
+    
+    # Ensure path starts with /
+    if not path.startswith("/"):
+        path = "/" + path
+    
+    return f"{base_url}{path}"
+
 # Note: config will be imported after parsing CLI args if --config is provided
 # For now, we'll delay the import and initialization
 app = FastAPI(title="AI Proxy Server")
 
-# Initialize Jinja2 templates
+# Add proxy headers middleware (must be added before other middleware)
+app.add_middleware(ProxyHeadersMiddleware)
+
+# Initialize Jinja2 templates with custom globals for proxy-aware URLs
 templates = Jinja2Templates(directory="templates")
 
-# Add session middleware (will be configured with secret key in main())
-# Placeholder - will be added in main() after we have a secret key
-session_secret_key = None
+# Add custom template globals for proxy-aware URL generation
+def setup_template_globals():
+    """Setup Jinja2 template globals for proxy-aware URLs"""
+    templates.env.globals['url_for'] = url_for
+    templates.env.globals['get_base_url'] = get_base_url
+
+# Call setup after templates are initialized
+setup_template_globals()
 
-# These will be initialized in main() after config is loaded
+# Add session middleware at module level with a temporary secret key
+# This is needed for uvicorn import (when main() doesn't run)
+_default_session_secret = secrets.token_urlsafe(32)
+app.add_middleware(SessionMiddleware, secret_key=_default_session_secret)
+
+# These will be initialized in startup event or main() after config is loaded
 request_handler = None
 rotation_handler = None
 autoselect_handler = None
 server_config = None
 config = None
+_initialized = False
+
+# Model cache for dynamically fetched provider models
+_model_cache = {}
+_model_cache_timestamps = {}
+_cache_refresh_interval = 4 * 3600  # 4 hours in seconds
+_cache_refresh_task = None
+
+def initialize_app(custom_config_dir=None):
+    """Initialize app globals. Called by startup event or main()."""
+    global config, request_handler, rotation_handler, autoselect_handler, server_config, _initialized
+    
+    if _initialized:
+        return
+    
+    # Set custom config directory if provided
+    if custom_config_dir:
+        set_config_dir(custom_config_dir)
+        logger.info(f"Using custom config directory: {custom_config_dir}")
+    
+    # Import config
+    from aisbf.config import config as cfg
+    from aisbf.handlers import RequestHandler, RotationHandler, AutoselectHandler
+    
+    config = cfg
+    request_handler = RequestHandler()
+    rotation_handler = RotationHandler()
+    autoselect_handler = AutoselectHandler()
+    
+    # Load server configuration
+    server_config = load_server_config(custom_config_dir)
+    
+    # Load dashboard config
+    aisbf_config_path = Path.home() / '.aisbf' / 'aisbf.json'
+    if not aisbf_config_path.exists():
+        if custom_config_dir:
+            aisbf_config_path = Path(custom_config_dir) / 'aisbf.json'
+        else:
+            # Try installed location first
+            installed_path = Path(__file__).parent / 'aisbf.json'
+            if installed_path.exists():
+                aisbf_config_path = installed_path
+            else:
+                # Fall back to config subdirectory
+                aisbf_config_path = Path(__file__).parent / 'config' / 'aisbf.json'
+    
+    if aisbf_config_path.exists():
+        with open(aisbf_config_path) as f:
+            aisbf_config = json.load(f)
+            server_config['dashboard_config'] = aisbf_config.get('dashboard', {})
+    else:
+        # Default with hashed password for 'admin'
+        server_config['dashboard_config'] = {
+            'username': 'admin', 
+            'password': '8c6976e5b5410415bde908bd4dee15dfb167a9c873fc4bb8a81f6f2ab448a918'
+        }
+    
+    _initialized = True
+    logger.info("App initialization complete")
+
+async def fetch_provider_models(provider_id: str) -> list:
+    """Fetch models from provider API and cache them"""
+    global _model_cache, _model_cache_timestamps
+    
+    try:
+        logger.debug(f"Fetching models from provider: {provider_id}")
+        # Create a dummy request object for the handler
+        from starlette.requests import Request
+        from starlette.datastructures import Headers
+        
+        scope = {
+            "type": "http",
+            "method": "GET",
+            "headers": [],
+            "query_string": b"",
+            "path": f"/api/{provider_id}/models",
+        }
+        dummy_request = Request(scope)
+        
+        # Fetch models from provider API
+        models = await request_handler.handle_model_list(dummy_request, provider_id)
+        
+        # Cache the results
+        _model_cache[provider_id] = models
+        _model_cache_timestamps[provider_id] = time.time()
+        
+        logger.info(f"Cached {len(models)} models from provider: {provider_id}")
+        return models
+    except Exception as e:
+        logger.warning(f"Failed to fetch models from provider {provider_id}: {e}")
+        return []
+
+async def refresh_model_cache():
+    """Background task to refresh model cache periodically"""
+    global _model_cache, _model_cache_timestamps
+    
+    while True:
+        try:
+            await asyncio.sleep(_cache_refresh_interval)
+            logger.info("Starting periodic model cache refresh...")
+            
+            # Refresh cache for all providers without local model config
+            for provider_id, provider_config in config.providers.items():
+                if not (hasattr(provider_config, 'models') and provider_config.models):
+                    await fetch_provider_models(provider_id)
+            
+            logger.info("Model cache refresh complete")
+        except Exception as e:
+            logger.error(f"Error in model cache refresh task: {e}")
+
+async def get_provider_models(provider_id: str, provider_config) -> list:
+    """Get models for a provider from local config or cache"""
+    global _model_cache, _model_cache_timestamps
+    
+    # Check if provider requires API key and if it's configured
+    api_key_required = getattr(provider_config, 'api_key_required', False)
+    api_key = getattr(provider_config, 'api_key', None)
+    
+    # If API key is required but not configured or is placeholder, skip this provider
+    if api_key_required:
+        if not api_key or api_key.startswith('YOUR_'):
+            logger.debug(f"Skipping provider {provider_id}: API key required but not configured")
+            return []
+    
+    # If provider has local model config, use it
+    if hasattr(provider_config, 'models') and provider_config.models:
+        models = []
+        for model in provider_config.models:
+            model_id = f"{provider_id}/{model.name}"
+            models.append({
+                'id': model_id,
+                'object': 'model',
+                'created': int(time.time()),
+                'owned_by': provider_config.name,
+                'provider': provider_id,
+                'type': 'provider',
+                'model_name': model.name,
+                'context_size': getattr(model, 'context_size', None),
+                'capabilities': getattr(model, 'capabilities', []),
+                'source': 'local_config'
+            })
+        return models
+    
+    # Check if we have cached models
+    if provider_id in _model_cache:
+        cache_age = time.time() - _model_cache_timestamps.get(provider_id, 0)
+        if cache_age < _cache_refresh_interval:
+            # Cache is still fresh, use it
+            cached_models = _model_cache[provider_id]
+            if cached_models:  # Only return if we have actual models
+                # Add provider prefix to model IDs
+                models = []
+                for model in cached_models:
+                    model_copy = model.copy()
+                    model_copy['id'] = f"{provider_id}/{model.get('id', model.get('name', ''))}"
+                    model_copy['provider'] = provider_id
+                    model_copy['type'] = 'provider'
+                    model_copy['source'] = 'api_cache'
+                    models.append(model_copy)
+                return models
+    
+    # No local config and no cache, try to fetch from API (only if API key is valid or not required)
+    if not api_key_required or (api_key and not api_key.startswith('YOUR_')):
+        try:
+            fetched_models = await fetch_provider_models(provider_id)
+            if fetched_models:
+                # Add provider prefix to model IDs
+                models = []
+                for model in fetched_models:
+                    model_copy = model.copy()
+                    model_copy['id'] = f"{provider_id}/{model.get('id', model.get('name', ''))}"
+                    model_copy['provider'] = provider_id
+                    model_copy['type'] = 'provider'
+                    model_copy['source'] = 'api_cache'
+                    models.append(model_copy)
+                return models
+        except Exception as e:
+            logger.debug(f"Failed to fetch models for provider {provider_id}: {e}")
+    
+    # No models available - return empty list (don't show generic fallback)
+    return []
+
+@app.on_event("startup")
+async def startup_event():
+    """Initialize app on startup (for uvicorn import case)."""
+    global config, server_config, _cache_refresh_task
+    if not _initialized:
+        # Use environment variable for config dir if set
+        custom_config_dir = get_config_dir()
+        initialize_app(custom_config_dir)
+    
+    # Log configuration files loaded
+    if config and hasattr(config, '_loaded_files'):
+        logger.info("")
+        logger.info("=" * 80)
+        logger.info("=== CONFIGURATION FILES LOADED ===")
+        logger.info("=" * 80)
+        
+        if 'providers' in config._loaded_files:
+            logger.info(f"Providers:    {config._loaded_files['providers']}")
+        
+        if 'rotations' in config._loaded_files:
+            logger.info(f"Rotations:    {config._loaded_files['rotations']}")
+        
+        if 'autoselect' in config._loaded_files:
+            logger.info(f"Autoselect:   {config._loaded_files['autoselect']}")
+        
+        if 'condensation' in config._loaded_files:
+            logger.info(f"Condensation: {config._loaded_files['condensation']}")
+        
+        logger.info("=" * 80)
+        logger.info("")
+    
+    # Start background task for model cache refresh
+    if _cache_refresh_task is None:
+        _cache_refresh_task = asyncio.create_task(refresh_model_cache())
+        logger.info(f"Started model cache refresh task (interval: {_cache_refresh_interval/3600} hours)")
+    
+    # In debug mode, validate provider configurations
+    AISBF_DEBUG = os.environ.get('AISBF_DEBUG', '').lower() in ('true', '1', 'yes')
+    if AISBF_DEBUG and config:
+        logger.info("")
+        logger.info("=" * 80)
+        logger.info("=== PROVIDER CONFIGURATION VALIDATION (DEBUG MODE) ===")
+        logger.info("=" * 80)
+        
+        for provider_id, provider_config in config.providers.items():
+            logger.info(f"")
+            logger.info(f"Provider: {provider_id}")
+            logger.info(f"  Type: {provider_config.type}")
+            logger.info(f"  Endpoint: {provider_config.endpoint}")
+            logger.info(f"  API Key Required: {provider_config.api_key_required}")
+            
+            # Check if API key is configured
+            if provider_config.api_key_required:
+                if provider_config.api_key:
+                    logger.info(f"  API Key: Configured ✓")
+                    logger.info(f"  Status: Ready to use")
+                else:
+                    logger.warning(f"  API Key: NOT CONFIGURED ✗")
+                    logger.warning(f"  Status: WILL BE SKIPPED - API key required but not provided")
+                    logger.warning(f"  Action: Add api_key to provider configuration in providers.json")
+            else:
+                logger.info(f"  API Key: Not required")
+                logger.info(f"  Status: Ready to use")
+            
+            # Show model count if available
+            if provider_config.models:
+                logger.info(f"  Models Configured: {len(provider_config.models)}")
+                for model in provider_config.models[:3]:  # Show first 3 models
+                    logger.info(f"    - {model.name}")
+                if len(provider_config.models) > 3:
+                    logger.info(f"    ... and {len(provider_config.models) - 3} more")
+            else:
+                logger.info(f"  Models Configured: None (will use provider's default models)")
+        
+        logger.info("")
+        logger.info("=" * 80)
+        logger.info("")
+    
+    logger.info(f"=== AISBF Server Started ===")
+    logger.info(f"Available providers: {list(config.providers.keys()) if config else []}")
+    logger.info(f"Available rotations: {list(config.rotations.keys()) if config else []}")
+    logger.info(f"Available autoselect: {list(config.autoselect.keys()) if config else []}")
 
 # Authentication middleware
 @app.middleware("http")
 async def auth_middleware(request: Request, call_next):
     """Check API token authentication if enabled"""
-    if server_config.get('auth_enabled', False):
+    if server_config and server_config.get('auth_enabled', False):
         # Skip auth for root endpoint and dashboard routes
         if request.url.path == "/" or request.url.path.startswith("/dashboard"):
             response = await call_next(request)
             return response
         
+        # Skip auth for public models endpoints (GET only)
+        if request.method == "GET" and request.url.path in ["/api/models", "/api/v1/models"]:
+            response = await call_next(request)
+            return response
+        
         # Check for Authorization header
         auth_header = request.headers.get('Authorization', '')
         
@@ -447,12 +831,12 @@ app.add_middleware(
 @app.get("/dashboard/login", response_class=HTMLResponse)
 async def dashboard_login_page(request: Request):
     """Show dashboard login page"""
-    return templates.TemplateResponse("dashboard/login.html", {"request": request})
+    return templates.TemplateResponse("dashboard/login.html", {"request": request, "session": request.session})
 
 @app.post("/dashboard/login")
 async def dashboard_login(request: Request, username: str = Form(...), password: str = Form(...)):
     """Handle dashboard login"""
-    dashboard_config = server_config.get('dashboard_config', {})
+    dashboard_config = server_config.get('dashboard_config', {}) if server_config else {}
     stored_username = dashboard_config.get('username', 'admin')
     stored_password_hash = dashboard_config.get('password', '8c6976e5b5410415bde908bd4dee15dfb167a9c873fc4bb8a81f6f2ab448a918')
     
@@ -463,19 +847,19 @@ async def dashboard_login(request: Request, username: str = Form(...), password:
     if username == stored_username and password_hash == stored_password_hash:
         request.session['logged_in'] = True
         request.session['username'] = username
-        return RedirectResponse(url="/dashboard", status_code=303)
-    return templates.TemplateResponse("dashboard/login.html", {"request": request, "error": "Invalid credentials"})
+        return RedirectResponse(url=url_for(request, "/dashboard"), status_code=303)
+    return templates.TemplateResponse("dashboard/login.html", {"request": request, "session": request.session, "error": "Invalid credentials"})
 
 @app.get("/dashboard/logout")
 async def dashboard_logout(request: Request):
     """Handle dashboard logout"""
     request.session.clear()
-    return RedirectResponse(url="/dashboard/login", status_code=303)
+    return RedirectResponse(url=url_for(request, "/dashboard/login"), status_code=303)
 
 def require_dashboard_auth(request: Request):
     """Check if user is logged in to dashboard"""
     if not request.session.get('logged_in'):
-        return RedirectResponse(url="/dashboard/login", status_code=303)
+        return RedirectResponse(url=url_for(request, "/dashboard/login"), status_code=303)
     return None
 
 @app.get("/dashboard", response_class=HTMLResponse)
@@ -507,13 +891,23 @@ async def dashboard_providers(request: Request):
         config_path = Path(__file__).parent / 'config' / 'providers.json'
     
     with open(config_path) as f:
-        config_content = f.read()
+        full_config = json.load(f)
     
-    return templates.TemplateResponse("dashboard/edit_config.html", {
+    # Extract just the providers object (handle both nested and flat structures)
+    if 'providers' in full_config and isinstance(full_config['providers'], dict):
+        providers_data = full_config['providers']
+    else:
+        # Fallback for flat structure (backward compatibility)
+        providers_data = {k: v for k, v in full_config.items() if k != 'condensation'}
+    
+    # Check for success parameter
+    success = request.query_params.get('success')
+    
+    return templates.TemplateResponse("dashboard/providers.html", {
         "request": request,
         "session": request.session,
-        "title": "Providers Configuration",
-        "config_content": config_content
+        "providers_json": json.dumps(providers_data),
+        "success": "Configuration saved successfully! Restart server for changes to take effect." if success else None
     })
 
 @app.post("/dashboard/providers")
@@ -525,27 +919,58 @@ async def dashboard_providers_save(request: Request, config: str = Form(...)):
     
     try:
         # Validate JSON
-        json.loads(config)
+        providers_data = json.loads(config)
+        
+        # Apply defaults: if condense_method is set but condense_context is not, default to 80
+        for provider_key, provider in providers_data.items():
+            if 'models' in provider and isinstance(provider['models'], list):
+                for model in provider['models']:
+                    if 'condense_method' in model and model.get('condense_method'):
+                        if 'condense_context' not in model or model.get('condense_context') is None:
+                            model['condense_context'] = 80
         
-        # Save to file
+        # Load existing config to preserve structure
         config_path = Path.home() / '.aisbf' / 'providers.json'
-        config_path.parent.mkdir(parents=True, exist_ok=True)
-        with open(config_path, 'w') as f:
-            f.write(config)
+        if not config_path.exists():
+            config_path = Path(__file__).parent / 'config' / 'providers.json'
+        
+        # Read existing config to preserve condensation settings
+        with open(config_path) as f:
+            full_config = json.load(f)
         
-        return templates.TemplateResponse("dashboard/edit_config.html", {
+        # Update providers section while preserving other keys
+        full_config['providers'] = providers_data
+        
+        # Save to file with full structure
+        save_path = Path.home() / '.aisbf' / 'providers.json'
+        save_path.parent.mkdir(parents=True, exist_ok=True)
+        with open(save_path, 'w') as f:
+            json.dump(full_config, f, indent=2)
+        
+        return templates.TemplateResponse("dashboard/providers.html", {
             "request": request,
             "session": request.session,
-            "title": "Providers Configuration",
-            "config_content": config,
+            "providers_json": json.dumps(providers_data),
             "success": "Configuration saved successfully! Restart server for changes to take effect."
         })
     except json.JSONDecodeError as e:
-        return templates.TemplateResponse("dashboard/edit_config.html", {
+        # Reload current config on error
+        config_path = Path.home() / '.aisbf' / 'providers.json'
+        if not config_path.exists():
+            config_path = Path(__file__).parent / 'config' / 'providers.json'
+        with open(config_path) as f:
+            full_config = json.load(f)
+        
+        # Extract providers
+        if 'providers' in full_config and isinstance(full_config['providers'], dict):
+            providers_data = full_config['providers']
+        else:
+            providers_data = {k: v for k, v in full_config.items() if k != 'condensation'}
+        
+        return templates.TemplateResponse("dashboard/providers.html", {
             "request": request,
             "session": request.session,
-            "title": "Providers Configuration",
-            "config_content": config,
+            "providers_json": json.dumps(providers_data),
             "error": f"Invalid JSON: {str(e)}"
         })
 
@@ -561,13 +986,20 @@ async def dashboard_rotations(request: Request):
         config_path = Path(__file__).parent / 'config' / 'rotations.json'
     
     with open(config_path) as f:
-        config_content = f.read()
+        rotations_data = json.load(f)
+    
+    # Get available providers
+    available_providers = list(config.providers.keys()) if config else []
     
-    return templates.TemplateResponse("dashboard/edit_config.html", {
+    # Check for success parameter
+    success = request.query_params.get('success')
+    
+    return templates.TemplateResponse("dashboard/rotations.html", {
         "request": request,
         "session": request.session,
-        "title": "Rotations Configuration",
-        "config_content": config_content
+        "rotations_json": json.dumps(rotations_data),
+        "available_providers": json.dumps(available_providers),
+        "success": "Configuration saved successfully! Restart server for changes to take effect." if success else None
     })
 
 @app.post("/dashboard/rotations")
@@ -578,25 +1010,48 @@ async def dashboard_rotations_save(request: Request, config: str = Form(...)):
         return auth_check
     
     try:
-        json.loads(config)
+        rotations_data = json.loads(config)
+        
+        # Apply defaults: if condense_method is set but condense_context is not, default to 80
+        if 'rotations' in rotations_data:
+            for rotation_key, rotation in rotations_data['rotations'].items():
+                if 'providers' in rotation and isinstance(rotation['providers'], list):
+                    for provider in rotation['providers']:
+                        if 'models' in provider and isinstance(provider['models'], list):
+                            for model in provider['models']:
+                                if 'condense_method' in model and model.get('condense_method'):
+                                    if 'condense_context' not in model or model.get('condense_context') is None:
+                                        model['condense_context'] = 80
+        
         config_path = Path.home() / '.aisbf' / 'rotations.json'
         config_path.parent.mkdir(parents=True, exist_ok=True)
         with open(config_path, 'w') as f:
-            f.write(config)
+            json.dump(rotations_data, f, indent=2)
         
-        return templates.TemplateResponse("dashboard/edit_config.html", {
+        available_providers = list(config.providers.keys()) if config else []
+        
+        return templates.TemplateResponse("dashboard/rotations.html", {
             "request": request,
             "session": request.session,
-            "title": "Rotations Configuration",
-            "config_content": config,
+            "rotations_json": json.dumps(rotations_data),
+            "available_providers": json.dumps(available_providers),
             "success": "Configuration saved successfully! Restart server for changes to take effect."
         })
     except json.JSONDecodeError as e:
-        return templates.TemplateResponse("dashboard/edit_config.html", {
+        # Reload current config on error
+        config_path = Path.home() / '.aisbf' / 'rotations.json'
+        if not config_path.exists():
+            config_path = Path(__file__).parent / 'config' / 'rotations.json'
+        with open(config_path) as f:
+            rotations_data = json.load(f)
+        
+        available_providers = list(config.providers.keys()) if config else []
+        
+        return templates.TemplateResponse("dashboard/rotations.html", {
             "request": request,
             "session": request.session,
-            "title": "Rotations Configuration",
-            "config_content": config,
+            "rotations_json": json.dumps(rotations_data),
+            "available_providers": json.dumps(available_providers),
             "error": f"Invalid JSON: {str(e)}"
         })
 
@@ -612,13 +1067,52 @@ async def dashboard_autoselect(request: Request):
         config_path = Path(__file__).parent / 'config' / 'autoselect.json'
     
     with open(config_path) as f:
-        config_content = f.read()
+        autoselect_data = json.load(f)
+    
+    # Get available rotations
+    available_rotations = list(config.rotations.keys()) if config else []
+    
+    # Get available provider models
+    available_models = []
+    
+    # Add rotation IDs
+    for rotation_id in available_rotations:
+        available_models.append({
+            'id': rotation_id,
+            'name': f'{rotation_id} (rotation)',
+            'type': 'rotation'
+        })
+    
+    # Add provider models
+    providers_path = Path.home() / '.aisbf' / 'providers.json'
+    if not providers_path.exists():
+        providers_path = Path(__file__).parent / 'config' / 'providers.json'
+    
+    if providers_path.exists():
+        with open(providers_path) as f:
+            providers_config = json.load(f)
+            providers_data = providers_config.get('providers', {})
+            
+            for provider_id, provider in providers_data.items():
+                if 'models' in provider and isinstance(provider['models'], list):
+                    for model in provider['models']:
+                        model_id = f"{provider_id}/{model['name']}"
+                        available_models.append({
+                            'id': model_id,
+                            'name': f"{model_id} (provider model)",
+                            'type': 'provider'
+                        })
+    
+    # Check for success parameter
+    success = request.query_params.get('success')
     
-    return templates.TemplateResponse("dashboard/edit_config.html", {
+    return templates.TemplateResponse("dashboard/autoselect.html", {
         "request": request,
         "session": request.session,
-        "title": "Autoselect Configuration",
-        "config_content": config_content
+        "autoselect_json": json.dumps(autoselect_data),
+        "available_rotations": json.dumps(available_rotations),
+        "available_models": json.dumps(available_models),
+        "success": "Configuration saved successfully! Restart server for changes to take effect." if success else None
     })
 
 @app.post("/dashboard/autoselect")
@@ -629,53 +1123,120 @@ async def dashboard_autoselect_save(request: Request, config: str = Form(...)):
         return auth_check
     
     try:
-        json.loads(config)
+        autoselect_data = json.loads(config)
         config_path = Path.home() / '.aisbf' / 'autoselect.json'
         config_path.parent.mkdir(parents=True, exist_ok=True)
         with open(config_path, 'w') as f:
-            f.write(config)
+            json.dump(autoselect_data, f, indent=2)
+        
+        available_rotations = list(config.rotations.keys()) if config else []
         
-        return templates.TemplateResponse("dashboard/edit_config.html", {
+        return templates.TemplateResponse("dashboard/autoselect.html", {
             "request": request,
             "session": request.session,
-            "title": "Autoselect Configuration",
-            "config_content": config,
+            "autoselect_json": json.dumps(autoselect_data),
+            "available_rotations": json.dumps(available_rotations),
             "success": "Configuration saved successfully! Restart server for changes to take effect."
         })
     except json.JSONDecodeError as e:
-        return templates.TemplateResponse("dashboard/edit_config.html", {
+        # Reload current config on error
+        config_path = Path.home() / '.aisbf' / 'autoselect.json'
+        if not config_path.exists():
+            config_path = Path(__file__).parent / 'config' / 'autoselect.json'
+        with open(config_path) as f:
+            autoselect_data = json.load(f)
+        
+        available_rotations = list(config.rotations.keys()) if config else []
+        
+        return templates.TemplateResponse("dashboard/autoselect.html", {
             "request": request,
             "session": request.session,
-            "title": "Autoselect Configuration",
-            "config_content": config,
+            "autoselect_json": json.dumps(autoselect_data),
+            "available_rotations": json.dumps(available_rotations),
             "error": f"Invalid JSON: {str(e)}"
         })
 
-@app.get("/dashboard/condensation", response_class=HTMLResponse)
-async def dashboard_condensation(request: Request):
-    """Edit condensation prompts"""
+@app.get("/dashboard/prompts", response_class=HTMLResponse)
+async def dashboard_prompts(request: Request):
+    """Edit prompt templates"""
     auth_check = require_dashboard_auth(request)
     if auth_check:
         return auth_check
     
-    # Load condensation_conversational.md
-    config_path = Path.home() / '.aisbf' / 'condensation_conversational.md'
+    # Define available prompts
+    prompt_files = [
+        {'key': 'condensation_conversational', 'name': 'Condensation - Conversational', 'filename': 'condensation_conversational.md'},
+        {'key': 'condensation_semantic', 'name': 'Condensation - Semantic', 'filename': 'condensation_semantic.md'},
+        {'key': 'autoselect', 'name': 'Autoselect - Model Selection', 'filename': 'autoselect.md'},
+    ]
+    
+    prompts_data = []
+    for prompt_file in prompt_files:
+        config_path = Path.home() / '.aisbf' / prompt_file['filename']
         if not config_path.exists():
-        config_path = Path(__file__).parent / 'config' / 'condensation_conversational.md'
+            config_path = Path(__file__).parent / 'config' / prompt_file['filename']
         
+        if config_path.exists():
             with open(config_path) as f:
-        config_content = f.read()
+                content = f.read()
+            prompts_data.append({
+                'key': prompt_file['key'],
+                'name': prompt_file['name'],
+                'filename': prompt_file['filename'],
+                'content': content
+            })
+    
+    # Check for success parameter
+    success = request.query_params.get('success')
     
-    return templates.TemplateResponse("dashboard/edit_config.html", {
+    return templates.TemplateResponse("dashboard/prompts.html", {
         "request": request,
         "session": request.session,
-        "title": "Condensation Prompts (Conversational)",
-        "config_content": config_content
+        "prompts": prompt_files,
+        "prompts_data": json.dumps(prompts_data),
+        "success": "Prompt saved successfully!" if success else None
     })
 
+@app.post("/dashboard/prompts")
+async def dashboard_prompts_save(request: Request, prompt_key: str = Form(...), prompt_content: str = Form(...)):
+    """Save prompt template"""
+    auth_check = require_dashboard_auth(request)
+    if auth_check:
+        return auth_check
+    
+    # Map prompt keys to filenames
+    prompt_map = {
+        'condensation_conversational': 'condensation_conversational.md',
+        'condensation_semantic': 'condensation_semantic.md',
+        'autoselect': 'autoselect.md',
+    }
+    
+    if prompt_key not in prompt_map:
+        return templates.TemplateResponse("dashboard/prompts.html", {
+            "request": request,
+            "session": request.session,
+            "prompts": [],
+            "prompts_data": "[]",
+            "error": "Invalid prompt key"
+        })
+    
+    filename = prompt_map[prompt_key]
+    config_path = Path.home() / '.aisbf' / filename
+    config_path.parent.mkdir(parents=True, exist_ok=True)
+    
+    with open(config_path, 'w') as f:
+        f.write(prompt_content)
+    
+    return RedirectResponse(url=url_for(request, "/dashboard/prompts?success=1"), status_code=303)
+
+@app.get("/dashboard/condensation", response_class=HTMLResponse)
+async def dashboard_condensation(request: Request):
+    """Redirect to prompts page for backward compatibility"""
+    return RedirectResponse(url=url_for(request, "/dashboard/prompts"), status_code=303)
+
 @app.post("/dashboard/condensation")
 async def dashboard_condensation_save(request: Request, config: str = Form(...)):
-    """Save condensation prompts"""
+    """Save condensation prompts - backward compatibility"""
     auth_check = require_dashboard_auth(request)
     if auth_check:
         return auth_check
@@ -685,13 +1246,7 @@ async def dashboard_condensation_save(request: Request, config: str = Form(...))
     with open(config_path, 'w') as f:
         f.write(config)
     
-    return templates.TemplateResponse("dashboard/edit_config.html", {
-        "request": request,
-        "session": request.session,
-        "title": "Condensation Prompts (Conversational)",
-        "config_content": config,
-        "success": "Prompts saved successfully!"
-    })
+    return RedirectResponse(url=url_for(request, "/dashboard/prompts?success=1"), status_code=303)
 
 @app.get("/dashboard/settings", response_class=HTMLResponse)
 async def dashboard_settings(request: Request):
@@ -724,7 +1279,8 @@ async def dashboard_settings_save(
     auth_tokens: str = Form(""),
     dashboard_username: str = Form(...),
     dashboard_password: str = Form(""),
-    internal_model_id: str = Form(...)
+    condensation_model_id: str = Form(...),
+    autoselect_model_id: str = Form(...)
 ):
     """Save server settings"""
     auth_check = require_dashboard_auth(request)
@@ -749,7 +1305,8 @@ async def dashboard_settings_save(
     if dashboard_password:  # Only update if provided - hash the password
         password_hash = hashlib.sha256(dashboard_password.encode()).hexdigest()
         aisbf_config['dashboard']['password'] = password_hash
-    aisbf_config['internal_model']['model_id'] = internal_model_id
+    aisbf_config['internal_model']['condensation_model_id'] = condensation_model_id
+    aisbf_config['internal_model']['autoselect_model_id'] = autoselect_model_id
     
     # Save config
     config_path = Path.home() / '.aisbf' / 'aisbf.json'
@@ -788,6 +1345,93 @@ async def dashboard_restart(request: Request):
     
     return JSONResponse({"message": "Server is restarting..."})
 
+@app.get("/dashboard/docs", response_class=HTMLResponse)
+async def dashboard_docs(request: Request):
+    """Display documentation"""
+    auth_check = require_dashboard_auth(request)
+    if auth_check:
+        return auth_check
+    
+    # Try to find DOCUMENTATION.md
+    doc_path = Path(__file__).parent / 'DOCUMENTATION.md'
+    if not doc_path.exists():
+        doc_path = Path.home() / '.aisbf' / 'DOCUMENTATION.md'
+    
+    if doc_path.exists():
+        with open(doc_path, encoding='utf-8') as f:
+            markdown_content = f.read()
+            # Convert markdown to HTML with extensions for better formatting
+            html_content = markdown.markdown(
+                markdown_content,
+                extensions=['fenced_code', 'tables', 'nl2br', 'sane_lists']
+            )
+    else:
+        html_content = "<p>Documentation file not found.</p>"
+    
+    return templates.TemplateResponse("dashboard/docs.html", {
+        "request": request,
+        "session": request.session,
+        "content": html_content,
+        "title": "Documentation"
+    })
+
+@app.get("/dashboard/about", response_class=HTMLResponse)
+async def dashboard_about(request: Request):
+    """Display README/About"""
+    auth_check = require_dashboard_auth(request)
+    if auth_check:
+        return auth_check
+    
+    # Try to find README.md
+    readme_path = Path(__file__).parent / 'README.md'
+    if not readme_path.exists():
+        readme_path = Path.home() / '.aisbf' / 'README.md'
+    
+    if readme_path.exists():
+        with open(readme_path, encoding='utf-8') as f:
+            markdown_content = f.read()
+            # Convert markdown to HTML with extensions for better formatting
+            html_content = markdown.markdown(
+                markdown_content,
+                extensions=['fenced_code', 'tables', 'nl2br', 'sane_lists']
+            )
+    else:
+        html_content = "<p>README file not found.</p>"
+    
+    return templates.TemplateResponse("dashboard/docs.html", {
+        "request": request,
+        "session": request.session,
+        "content": html_content,
+        "title": "About"
+    })
+
+@app.get("/dashboard/license", response_class=HTMLResponse)
+async def dashboard_license(request: Request):
+    """Display License"""
+    auth_check = require_dashboard_auth(request)
+    if auth_check:
+        return auth_check
+    
+    # Try to find LICENSE.txt
+    license_path = Path(__file__).parent / 'LICENSE.txt'
+    if not license_path.exists():
+        license_path = Path.home() / '.aisbf' / 'LICENSE.txt'
+    
+    if license_path.exists():
+        with open(license_path, encoding='utf-8') as f:
+            content = f.read()
+            # Convert to HTML with pre tags to preserve formatting
+            html_content = f"<pre style='white-space: pre-wrap; word-wrap: break-word; background: #0f3460; padding: 20px; border-radius: 6px; color: #e0e0e0; font-family: inherit;'>{content}</pre>"
+    else:
+        html_content = "<p>License file not found.</p>"
+    
+    return templates.TemplateResponse("dashboard/docs.html", {
+        "request": request,
+        "session": request.session,
+        "content": html_content,
+        "title": "License"
+    })
+
 def parse_provider_from_model(model: str) -> tuple[str, str]:
     """
     Parse provider and model from model field.
@@ -814,6 +1458,58 @@ async def root():
         "autoselect": list(config.autoselect.keys())
     }
 
+
+@app.get("/health")
+async def health():
+    return {"status": "ok"}
+
+
+@app.get("/api/v1/models/{model_id}")
+async def v1_get_model(model_id: str, request: Request):
+    """Get a specific model by ID (OpenAI-compatible endpoint)"""
+    # First try to find in all models
+    all_models_response = await v1_list_all_models(request)
+    all_models = all_models_response.get("data", [])
+    
+    for model in all_models:
+        if model.get("id") == model_id:
+            return model
+    
+    from fastapi import HTTPException
+    raise HTTPException(status_code=404, detail=f"Model '{model_id}' not found")
+
+
+@app.post("/api/v1/completions")
+async def v1_completions(request: Request):
+    """
+    Legacy text completions endpoint (OpenAI-compatible)
+    Maps to chat/completions for compatibility
+    """
+    # Get the request body
+    body = await request.body()
+    import json
+    data = json.loads(body) if body else {}
+    
+    # Convert completion request to chat completion
+    prompt = data.get("prompt", "")
+    model = data.get("model", "")
+    max_tokens = data.get("max_tokens", 2048)
+    temperature = data.get("temperature", 1.0)
+    
+    # Build chat messages from prompt
+    messages = [{"role": "user", "content": prompt}]
+    
+    # Create a new request for chat/completions
+    chat_request = {
+        "model": model,
+        "messages": messages,
+        "max_tokens": max_tokens,
+        "temperature": temperature
+    }
+    
+    # Call chat completions handler
+    return await v1_chat_completions(chat_request, request)
+
 # Standard OpenAI-compatible v1 endpoints
 @app.post("/api/v1/chat/completions")
 async def v1_chat_completions(request: Request, body: ChatCompletionRequest):
@@ -827,85 +1523,152 @@ async def v1_chat_completions(request: Request, body: ChatCompletionRequest):
     if not provider_id:
         raise HTTPException(
             status_code=400,
-            detail="Model must be in format 'provider/model' (e.g., 'openai/gpt-4')"
+            detail="Model must be in format 'provider/model', 'rotation/name', or 'autoselect/name'"
         )
     
     logger.info(f"Parsed provider: {provider_id}, model: {actual_model}")
     
     # Update body with actual model name
     body_dict = body.model_dump()
-    body_dict['model'] = actual_model
     
-    # Check if it's an autoselect
-    if provider_id in config.autoselect:
+    # PATH 3: Check if it's an autoselect (format: autoselect/{name})
+    if provider_id == "autoselect":
+        if actual_model not in config.autoselect:
+            raise HTTPException(
+                status_code=400,
+                detail=f"Autoselect '{actual_model}' not found. Available: {list(config.autoselect.keys())}"
+            )
+        body_dict['model'] = actual_model
         if body.stream:
-            return await autoselect_handler.handle_autoselect_streaming_request(provider_id, body_dict)
+            return await autoselect_handler.handle_autoselect_streaming_request(actual_model, body_dict)
         else:
-            return await autoselect_handler.handle_autoselect_request(provider_id, body_dict)
+            return await autoselect_handler.handle_autoselect_request(actual_model, body_dict)
     
-    # Check if it's a rotation
-    if provider_id in config.rotations:
-        return await rotation_handler.handle_rotation_request(provider_id, body_dict)
+    # PATH 2: Check if it's a rotation (format: rotation/{name})
+    if provider_id == "rotation":
+        if actual_model not in config.rotations:
+            raise HTTPException(
+                status_code=400,
+                detail=f"Rotation '{actual_model}' not found. Available: {list(config.rotations.keys())}"
+            )
+        body_dict['model'] = actual_model
+        return await rotation_handler.handle_rotation_request(actual_model, body_dict)
     
-    # Check if it's a provider
+    # PATH 1: Direct provider model (format: {provider}/{model})
     if provider_id not in config.providers:
         raise HTTPException(
             status_code=400,
-            detail=f"Provider '{provider_id}' not found. Available: {list(config.providers.keys())}"
+            detail=f"Provider '{provider_id}' not found. Available providers: {list(config.providers.keys())}, or use 'rotation/name' or 'autoselect/name'"
         )
     
     # Handle as direct provider request
+    body_dict['model'] = actual_model
     if body.stream:
         return await request_handler.handle_streaming_chat_completion(request, provider_id, body_dict)
     else:
         return await request_handler.handle_chat_completion(request, provider_id, body_dict)
 
-@app.get("/api/v1/models")
-async def v1_list_all_models(request: Request):
-    """List all available models from all providers"""
-    logger.info("=== V1 LIST ALL MODELS REQUEST ===")
+@app.get("/api/models")
+async def list_all_models(request: Request):
+    """List all available models from all providers (public endpoint)"""
+    logger.info("=== LIST ALL MODELS REQUEST ===")
     
     all_models = []
     
-    # Add provider models
-    for provider_id in config.providers.keys():
+    # PATH 1: Add provider models (from local config or cached API results)
+    for provider_id, provider_config in config.providers.items():
         try:
-            models = await request_handler.handle_model_list(request, provider_id)
-            for model in models:
-                # Prepend provider to model ID
-                model['id'] = f"{provider_id}/{model.get('id', model.get('name', ''))}"
-                model['provider'] = provider_id
-                all_models.append(model)
+            provider_models = await get_provider_models(provider_id, provider_config)
+            all_models.extend(provider_models)
         except Exception as e:
             logger.warning(f"Error listing models for provider {provider_id}: {e}")
     
-    # Add rotation models
-    for rotation_id in config.rotations.keys():
+    # PATH 2: Add rotations as rotation/{rotation_name}
+    for rotation_id, rotation_config in config.rotations.items():
         try:
-            models = await rotation_handler.handle_rotation_model_list(rotation_id)
-            for model in models:
-                model['id'] = f"{rotation_id}/{model.get('name', '')}"
-                model['type'] = 'rotation'
-                all_models.append(model)
+            all_models.append({
+                'id': f"rotation/{rotation_id}",
+                'object': 'model',
+                'created': int(time.time()),
+                'owned_by': 'aisbf-rotation',
+                'type': 'rotation',
+                'rotation_id': rotation_id,
+                'model_name': rotation_config.model_name
+            })
         except Exception as e:
-            logger.warning(f"Error listing models for rotation {rotation_id}: {e}")
+            logger.warning(f"Error listing rotation {rotation_id}: {e}")
     
-    # Add autoselect models
-    for autoselect_id in config.autoselect.keys():
+    # PATH 3: Add autoselect as autoselect/{autoselect_name}
+    for autoselect_id, autoselect_config in config.autoselect.items():
         try:
-            models = await autoselect_handler.handle_autoselect_model_list(autoselect_id)
-            for model in models:
-                model['id'] = f"{autoselect_id}/{model.get('name', model.get('id', ''))}"
-                model['type'] = 'autoselect'
-                all_models.append(model)
+            all_models.append({
+                'id': f"autoselect/{autoselect_id}",
+                'object': 'model',
+                'created': int(time.time()),
+                'owned_by': 'aisbf-autoselect',
+                'type': 'autoselect',
+                'autoselect_id': autoselect_id,
+                'model_name': autoselect_config.model_name,
+                'description': autoselect_config.description
+            })
         except Exception as e:
-            logger.warning(f"Error listing models for autoselect {autoselect_id}: {e}")
+            logger.warning(f"Error listing autoselect {autoselect_id}: {e}")
     
+    logger.info(f"Returning {len(all_models)} total models")
     return {"object": "list", "data": all_models}
 
-@app.post("/api/v1/audio/transcriptions")
-async def v1_audio_transcriptions(request: Request):
-    """Standard audio transcription endpoint"""
+@app.get("/api/v1/models")
+async def v1_list_all_models(request: Request):
+    """List all available models from all providers (OpenAI-compatible endpoint)"""
+    logger.info("=== V1 LIST ALL MODELS REQUEST ===")
+    
+    all_models = []
+    
+    # PATH 1: Add provider models (from local config or cached API results)
+    for provider_id, provider_config in config.providers.items():
+        try:
+            provider_models = await get_provider_models(provider_id, provider_config)
+            all_models.extend(provider_models)
+        except Exception as e:
+            logger.warning(f"Error listing models for provider {provider_id}: {e}")
+    
+    # PATH 2: Add rotations as rotation/{rotation_name}
+    for rotation_id, rotation_config in config.rotations.items():
+        try:
+            all_models.append({
+                'id': f"rotation/{rotation_id}",
+                'object': 'model',
+                'created': int(time.time()),
+                'owned_by': 'aisbf-rotation',
+                'type': 'rotation',
+                'rotation_id': rotation_id,
+                'model_name': rotation_config.model_name
+            })
+        except Exception as e:
+            logger.warning(f"Error listing rotation {rotation_id}: {e}")
+    
+    # PATH 3: Add autoselect as autoselect/{autoselect_name}
+    for autoselect_id, autoselect_config in config.autoselect.items():
+        try:
+            all_models.append({
+                'id': f"autoselect/{autoselect_id}",
+                'object': 'model',
+                'created': int(time.time()),
+                'owned_by': 'aisbf-autoselect',
+                'type': 'autoselect',
+                'autoselect_id': autoselect_id,
+                'model_name': autoselect_config.model_name,
+                'description': autoselect_config.description
+            })
+        except Exception as e:
+            logger.warning(f"Error listing autoselect {autoselect_id}: {e}")
+    
+    logger.info(f"Returning {len(all_models)} total models")
+    return {"object": "list", "data": all_models}
+
+@app.post("/api/v1/audio/transcriptions")
+async def v1_audio_transcriptions(request: Request):
+    """Standard audio transcription endpoint (supports all three proxy paths)"""
     logger.info("=== V1 AUDIO TRANSCRIPTION REQUEST ===")
     
     form = await request.form()
@@ -916,7 +1679,46 @@ async def v1_audio_transcriptions(request: Request):
     if not provider_id:
         raise HTTPException(
             status_code=400,
-            detail="Model must be in format 'provider/model' (e.g., 'openai/whisper-1')"
+            detail="Model must be in format 'provider/model', 'rotation/name', or 'autoselect/name'"
+        )
+    
+    # Handle rotation
+    if provider_id == "rotation":
+        if actual_model not in config.rotations:
+            raise HTTPException(
+                status_code=400,
+                detail=f"Rotation '{actual_model}' not found. Available: {list(config.rotations.keys())}"
+            )
+        selected_provider, selected_model = rotation_handler._select_provider_and_model(actual_model)
+        provider_id = selected_provider
+        actual_model = selected_model
+    
+    # Handle autoselect
+    elif provider_id == "autoselect":
+        if actual_model not in config.autoselect:
+            raise HTTPException(
+                status_code=400,
+                detail=f"Autoselect '{actual_model}' not found. Available: {list(config.autoselect.keys())}"
+            )
+        autoselect_config = config.autoselect[actual_model]
+        fallback = autoselect_config.fallback
+        if '/' in fallback:
+            provider_id, actual_model = fallback.split('/', 1)
+        else:
+            if fallback in config.rotations:
+                selected_provider, selected_model = rotation_handler._select_provider_and_model(fallback)
+                provider_id = selected_provider
+                actual_model = selected_model
+            else:
+                raise HTTPException(
+                    status_code=400,
+                    detail=f"Invalid fallback configuration for autoselect '{actual_model}'"
+                )
+    
+    if provider_id not in config.providers:
+        raise HTTPException(
+            status_code=400,
+            detail=f"Provider '{provider_id}' not found. Available: {list(config.providers.keys())}"
         )
     
     # Create new form data with updated model
@@ -932,7 +1734,7 @@ async def v1_audio_transcriptions(request: Request):
 
 @app.post("/api/v1/audio/speech")
 async def v1_audio_speech(request: Request, body: dict):
-    """Standard text-to-speech endpoint"""
+    """Standard text-to-speech endpoint (supports all three proxy paths)"""
     logger.info("=== V1 TEXT-TO-SPEECH REQUEST ===")
     
     model = body.get('model', '')
@@ -941,7 +1743,46 @@ async def v1_audio_speech(request: Request, body: dict):
     if not provider_id:
         raise HTTPException(
             status_code=400,
-            detail="Model must be in format 'provider/model' (e.g., 'openai/tts-1')"
+            detail="Model must be in format 'provider/model', 'rotation/name', or 'autoselect/name'"
+        )
+    
+    # Handle rotation
+    if provider_id == "rotation":
+        if actual_model not in config.rotations:
+            raise HTTPException(
+                status_code=400,
+                detail=f"Rotation '{actual_model}' not found. Available: {list(config.rotations.keys())}"
+            )
+        selected_provider, selected_model = rotation_handler._select_provider_and_model(actual_model)
+        provider_id = selected_provider
+        actual_model = selected_model
+    
+    # Handle autoselect
+    elif provider_id == "autoselect":
+        if actual_model not in config.autoselect:
+            raise HTTPException(
+                status_code=400,
+                detail=f"Autoselect '{actual_model}' not found. Available: {list(config.autoselect.keys())}"
+            )
+        autoselect_config = config.autoselect[actual_model]
+        fallback = autoselect_config.fallback
+        if '/' in fallback:
+            provider_id, actual_model = fallback.split('/', 1)
+        else:
+            if fallback in config.rotations:
+                selected_provider, selected_model = rotation_handler._select_provider_and_model(fallback)
+                provider_id = selected_provider
+                actual_model = selected_model
+            else:
+                raise HTTPException(
+                    status_code=400,
+                    detail=f"Invalid fallback configuration for autoselect '{actual_model}'"
+                )
+    
+    if provider_id not in config.providers:
+        raise HTTPException(
+            status_code=400,
+            detail=f"Provider '{provider_id}' not found. Available: {list(config.providers.keys())}"
         )
     
     body['model'] = actual_model
@@ -949,7 +1790,7 @@ async def v1_audio_speech(request: Request, body: dict):
 
 @app.post("/api/v1/images/generations")
 async def v1_image_generations(request: Request, body: dict):
-    """Standard image generation endpoint"""
+    """Standard image generation endpoint (supports all three proxy paths)"""
     logger.info("=== V1 IMAGE GENERATION REQUEST ===")
     
     model = body.get('model', '')
@@ -958,7 +1799,46 @@ async def v1_image_generations(request: Request, body: dict):
     if not provider_id:
         raise HTTPException(
             status_code=400,
-            detail="Model must be in format 'provider/model' (e.g., 'openai/dall-e-3')"
+            detail="Model must be in format 'provider/model', 'rotation/name', or 'autoselect/name'"
+        )
+    
+    # Handle rotation
+    if provider_id == "rotation":
+        if actual_model not in config.rotations:
+            raise HTTPException(
+                status_code=400,
+                detail=f"Rotation '{actual_model}' not found. Available: {list(config.rotations.keys())}"
+            )
+        selected_provider, selected_model = rotation_handler._select_provider_and_model(actual_model)
+        provider_id = selected_provider
+        actual_model = selected_model
+    
+    # Handle autoselect
+    elif provider_id == "autoselect":
+        if actual_model not in config.autoselect:
+            raise HTTPException(
+                status_code=400,
+                detail=f"Autoselect '{actual_model}' not found. Available: {list(config.autoselect.keys())}"
+            )
+        autoselect_config = config.autoselect[actual_model]
+        fallback = autoselect_config.fallback
+        if '/' in fallback:
+            provider_id, actual_model = fallback.split('/', 1)
+        else:
+            if fallback in config.rotations:
+                selected_provider, selected_model = rotation_handler._select_provider_and_model(fallback)
+                provider_id = selected_provider
+                actual_model = selected_model
+            else:
+                raise HTTPException(
+                    status_code=400,
+                    detail=f"Invalid fallback configuration for autoselect '{actual_model}'"
+                )
+    
+    if provider_id not in config.providers:
+        raise HTTPException(
+            status_code=400,
+            detail=f"Provider '{provider_id}' not found. Available: {list(config.providers.keys())}"
         )
     
     body['model'] = actual_model
@@ -966,7 +1846,7 @@ async def v1_image_generations(request: Request, body: dict):
 
 @app.post("/api/v1/embeddings")
 async def v1_embeddings(request: Request, body: dict):
-    """Standard embeddings endpoint"""
+    """Standard embeddings endpoint (supports all three proxy paths)"""
     logger.info("=== V1 EMBEDDINGS REQUEST ===")
     
     model = body.get('model', '')
@@ -975,7 +1855,46 @@ async def v1_embeddings(request: Request, body: dict):
     if not provider_id:
         raise HTTPException(
             status_code=400,
-            detail="Model must be in format 'provider/model' (e.g., 'openai/text-embedding-ada-002')"
+            detail="Model must be in format 'provider/model', 'rotation/name', or 'autoselect/name'"
+        )
+    
+    # Handle rotation
+    if provider_id == "rotation":
+        if actual_model not in config.rotations:
+            raise HTTPException(
+                status_code=400,
+                detail=f"Rotation '{actual_model}' not found. Available: {list(config.rotations.keys())}"
+            )
+        selected_provider, selected_model = rotation_handler._select_provider_and_model(actual_model)
+        provider_id = selected_provider
+        actual_model = selected_model
+    
+    # Handle autoselect
+    elif provider_id == "autoselect":
+        if actual_model not in config.autoselect:
+            raise HTTPException(
+                status_code=400,
+                detail=f"Autoselect '{actual_model}' not found. Available: {list(config.autoselect.keys())}"
+            )
+        autoselect_config = config.autoselect[actual_model]
+        fallback = autoselect_config.fallback
+        if '/' in fallback:
+            provider_id, actual_model = fallback.split('/', 1)
+        else:
+            if fallback in config.rotations:
+                selected_provider, selected_model = rotation_handler._select_provider_and_model(fallback)
+                provider_id = selected_provider
+                actual_model = selected_model
+            else:
+                raise HTTPException(
+                    status_code=400,
+                    detail=f"Invalid fallback configuration for autoselect '{actual_model}'"
+                )
+    
+    if provider_id not in config.providers:
+        raise HTTPException(
+            status_code=400,
+            detail=f"Provider '{provider_id}' not found. Available: {list(config.providers.keys())}"
         )
     
     body['model'] = actual_model
@@ -1254,63 +2173,247 @@ async def list_models(request: Request, provider_id: str):
         logger.error(f"Error handling list_models: {str(e)}", exc_info=True)
         raise
 
-# Audio endpoints
-@app.post("/api/{provider_id}/audio/transcriptions")
-async def audio_transcriptions(provider_id: str, request: Request):
-    """Handle audio transcription requests"""
-    logger.info(f"=== AUDIO TRANSCRIPTION REQUEST ===")
-    logger.info(f"Provider ID: {provider_id}")
+# Audio endpoints (model specified in request as provider/model, rotation/name, or autoselect/name)
+@app.post("/api/audio/transcriptions")
+async def audio_transcriptions(request: Request):
+    """Handle audio transcription requests (supports all three proxy paths)"""
+    logger.info("=== AUDIO TRANSCRIPTION REQUEST ===")
     
-    # Get form data (audio file upload)
     form = await request.form()
+    model = form.get('model', '')
     
-    try:
-        result = await request_handler.handle_audio_transcription(request, provider_id, form)
-        return result
-    except Exception as e:
-        logger.error(f"Error handling audio transcription: {str(e)}", exc_info=True)
-        raise HTTPException(status_code=500, detail=str(e))
+    provider_id, actual_model = parse_provider_from_model(model)
     
-@app.post("/api/{provider_id}/audio/speech")
-async def audio_speech(provider_id: str, request: Request, body: dict):
-    """Handle text-to-speech requests"""
-    logger.info(f"=== TEXT-TO-SPEECH REQUEST ===")
-    logger.info(f"Provider ID: {provider_id}")
+    if not provider_id:
+        raise HTTPException(
+            status_code=400,
+            detail="Model must be in format 'provider/model', 'rotation/name', or 'autoselect/name'"
+        )
     
-    try:
-        result = await request_handler.handle_text_to_speech(request, provider_id, body)
-        return result
-    except Exception as e:
-        logger.error(f"Error handling text-to-speech: {str(e)}", exc_info=True)
-        raise HTTPException(status_code=500, detail=str(e))
+    # Handle rotation
+    if provider_id == "rotation":
+        if actual_model not in config.rotations:
+            raise HTTPException(
+                status_code=400,
+                detail=f"Rotation '{actual_model}' not found. Available: {list(config.rotations.keys())}"
+            )
+        # Select a provider from the rotation using weighted random selection
+        selected_provider, selected_model = rotation_handler._select_provider_and_model(actual_model)
+        provider_id = selected_provider
+        actual_model = selected_model
+    
+    # Handle autoselect
+    elif provider_id == "autoselect":
+        if actual_model not in config.autoselect:
+            raise HTTPException(
+                status_code=400,
+                detail=f"Autoselect '{actual_model}' not found. Available: {list(config.autoselect.keys())}"
+            )
+        # Use the fallback model from autoselect config
+        autoselect_config = config.autoselect[actual_model]
+        fallback = autoselect_config.fallback
+        # Parse the fallback to get provider and model
+        if '/' in fallback:
+            provider_id, actual_model = fallback.split('/', 1)
+        else:
+            # Fallback is a rotation, select from it
+            if fallback in config.rotations:
+                selected_provider, selected_model = rotation_handler._select_provider_and_model(fallback)
+                provider_id = selected_provider
+                actual_model = selected_model
+            else:
+                raise HTTPException(
+                    status_code=400,
+                    detail=f"Invalid fallback configuration for autoselect '{actual_model}'"
+                )
     
-# Image endpoints
-@app.post("/api/{provider_id}/images/generations")
-async def image_generations(provider_id: str, request: Request, body: dict):
-    """Handle image generation requests"""
-    logger.info(f"=== IMAGE GENERATION REQUEST ===")
-    logger.info(f"Provider ID: {provider_id}")
+    # Validate provider exists
+    if provider_id not in config.providers:
+        raise HTTPException(
+            status_code=400,
+            detail=f"Provider '{provider_id}' not found. Available: {list(config.providers.keys())}"
+        )
     
-    try:
-        result = await request_handler.handle_image_generation(request, provider_id, body)
-        return result
-    except Exception as e:
-        logger.error(f"Error handling image generation: {str(e)}", exc_info=True)
-        raise HTTPException(status_code=500, detail=str(e))
+    # Create new form data with updated model
+    from starlette.datastructures import FormData
+    updated_form = FormData()
+    for key, value in form.items():
+        if key == 'model':
+            updated_form[key] = actual_model
+        else:
+            updated_form[key] = value
     
-# Embeddings endpoint
-@app.post("/api/{provider_id}/embeddings")
-async def embeddings(provider_id: str, request: Request, body: dict):
-    """Handle embeddings requests"""
-    logger.info(f"=== EMBEDDINGS REQUEST ===")
-    logger.info(f"Provider ID: {provider_id}")
+    return await request_handler.handle_audio_transcription(request, provider_id, updated_form)
 
-    try:
-        result = await request_handler.handle_embeddings(request, provider_id, body)
-        return result
-    except Exception as e:
-        logger.error(f"Error handling embeddings: {str(e)}", exc_info=True)
-        raise HTTPException(status_code=500, detail=str(e))
+@app.post("/api/audio/speech")
+async def audio_speech(request: Request, body: dict):
+    """Handle text-to-speech requests (supports all three proxy paths)"""
+    logger.info("=== TEXT-TO-SPEECH REQUEST ===")
+    
+    model = body.get('model', '')
+    provider_id, actual_model = parse_provider_from_model(model)
+    
+    if not provider_id:
+        raise HTTPException(
+            status_code=400,
+            detail="Model must be in format 'provider/model', 'rotation/name', or 'autoselect/name'"
+        )
+    
+    # Handle rotation
+    if provider_id == "rotation":
+        if actual_model not in config.rotations:
+            raise HTTPException(
+                status_code=400,
+                detail=f"Rotation '{actual_model}' not found. Available: {list(config.rotations.keys())}"
+            )
+        selected_provider, selected_model = rotation_handler._select_provider_and_model(actual_model)
+        provider_id = selected_provider
+        actual_model = selected_model
+    
+    # Handle autoselect
+    elif provider_id == "autoselect":
+        if actual_model not in config.autoselect:
+            raise HTTPException(
+                status_code=400,
+                detail=f"Autoselect '{actual_model}' not found. Available: {list(config.autoselect.keys())}"
+            )
+        autoselect_config = config.autoselect[actual_model]
+        fallback = autoselect_config.fallback
+        if '/' in fallback:
+            provider_id, actual_model = fallback.split('/', 1)
+        else:
+            if fallback in config.rotations:
+                selected_provider, selected_model = rotation_handler._select_provider_and_model(fallback)
+                provider_id = selected_provider
+                actual_model = selected_model
+            else:
+                raise HTTPException(
+                    status_code=400,
+                    detail=f"Invalid fallback configuration for autoselect '{actual_model}'"
+                )
+    
+    if provider_id not in config.providers:
+        raise HTTPException(
+            status_code=400,
+            detail=f"Provider '{provider_id}' not found. Available: {list(config.providers.keys())}"
+        )
+    
+    body['model'] = actual_model
+    return await request_handler.handle_text_to_speech(request, provider_id, body)
+
+# Image endpoints (supports all three proxy paths)
+@app.post("/api/images/generations")
+async def image_generations(request: Request, body: dict):
+    """Handle image generation requests (supports all three proxy paths)"""
+    logger.info("=== IMAGE GENERATION REQUEST ===")
+    
+    model = body.get('model', '')
+    provider_id, actual_model = parse_provider_from_model(model)
+    
+    if not provider_id:
+        raise HTTPException(
+            status_code=400,
+            detail="Model must be in format 'provider/model', 'rotation/name', or 'autoselect/name'"
+        )
+    
+    # Handle rotation
+    if provider_id == "rotation":
+        if actual_model not in config.rotations:
+            raise HTTPException(
+                status_code=400,
+                detail=f"Rotation '{actual_model}' not found. Available: {list(config.rotations.keys())}"
+            )
+        selected_provider, selected_model = rotation_handler._select_provider_and_model(actual_model)
+        provider_id = selected_provider
+        actual_model = selected_model
+    
+    # Handle autoselect
+    elif provider_id == "autoselect":
+        if actual_model not in config.autoselect:
+            raise HTTPException(
+                status_code=400,
+                detail=f"Autoselect '{actual_model}' not found. Available: {list(config.autoselect.keys())}"
+            )
+        autoselect_config = config.autoselect[actual_model]
+        fallback = autoselect_config.fallback
+        if '/' in fallback:
+            provider_id, actual_model = fallback.split('/', 1)
+        else:
+            if fallback in config.rotations:
+                selected_provider, selected_model = rotation_handler._select_provider_and_model(fallback)
+                provider_id = selected_provider
+                actual_model = selected_model
+            else:
+                raise HTTPException(
+                    status_code=400,
+                    detail=f"Invalid fallback configuration for autoselect '{actual_model}'"
+                )
+    
+    if provider_id not in config.providers:
+        raise HTTPException(
+            status_code=400,
+            detail=f"Provider '{provider_id}' not found. Available: {list(config.providers.keys())}"
+        )
+    
+    body['model'] = actual_model
+    return await request_handler.handle_image_generation(request, provider_id, body)
+
+# Embeddings endpoint (supports all three proxy paths)
+@app.post("/api/embeddings")
+async def embeddings(request: Request, body: dict):
+    """Handle embeddings requests (supports all three proxy paths)"""
+    logger.info("=== EMBEDDINGS REQUEST ===")
+    
+    model = body.get('model', '')
+    provider_id, actual_model = parse_provider_from_model(model)
+    
+    if not provider_id:
+        raise HTTPException(
+            status_code=400,
+            detail="Model must be in format 'provider/model', 'rotation/name', or 'autoselect/name'"
+        )
+    
+    # Handle rotation
+    if provider_id == "rotation":
+        if actual_model not in config.rotations:
+            raise HTTPException(
+                status_code=400,
+                detail=f"Rotation '{actual_model}' not found. Available: {list(config.rotations.keys())}"
+            )
+        selected_provider, selected_model = rotation_handler._select_provider_and_model(actual_model)
+        provider_id = selected_provider
+        actual_model = selected_model
+    
+    # Handle autoselect
+    elif provider_id == "autoselect":
+        if actual_model not in config.autoselect:
+            raise HTTPException(
+                status_code=400,
+                detail=f"Autoselect '{actual_model}' not found. Available: {list(config.autoselect.keys())}"
+            )
+        autoselect_config = config.autoselect[actual_model]
+        fallback = autoselect_config.fallback
+        if '/' in fallback:
+            provider_id, actual_model = fallback.split('/', 1)
+        else:
+            if fallback in config.rotations:
+                selected_provider, selected_model = rotation_handler._select_provider_and_model(fallback)
+                provider_id = selected_provider
+                actual_model = selected_model
+            else:
+                raise HTTPException(
+                    status_code=400,
+                    detail=f"Invalid fallback configuration for autoselect '{actual_model}'"
+                )
+    
+    if provider_id not in config.providers:
+        raise HTTPException(
+            status_code=400,
+            detail=f"Provider '{provider_id}' not found. Available: {list(config.providers.keys())}"
+        )
+    
+    body['model'] = actual_model
+    return await request_handler.handle_embeddings(request, provider_id, body)
 
 # Content proxy endpoint
 @app.get("/api/proxy/{content_id}")
@@ -1382,42 +2485,11 @@ Examples:
     global _original_argv
     _original_argv = sys.argv.copy()
     
-    # Set custom config directory if provided
-    if args.config:
-        set_config_dir(args.config)
-        logger.info(f"Using custom config directory: {args.config}")
-    
-    # Import config after setting custom directory
-    global config, request_handler, rotation_handler, autoselect_handler, server_config
-    from aisbf.config import config
-    from aisbf.handlers import RequestHandler, RotationHandler, AutoselectHandler
+    # Initialize app (this sets config, handlers, server_config)
+    initialize_app(args.config)
     
-    # Initialize handlers
-    request_handler = RequestHandler()
-    rotation_handler = RotationHandler()
-    autoselect_handler = AutoselectHandler()
-    
-    # Load server configuration
-    server_config = load_server_config(args.config)
-    
-    # Add session middleware for dashboard
-    secret_key = secrets.token_urlsafe(32)
-    app.add_middleware(SessionMiddleware, secret_key=secret_key)
-    
-    # Load dashboard config
-    aisbf_config_path = Path.home() / '.aisbf' / 'aisbf.json'
-    if not aisbf_config_path.exists():
-        if args.config:
-            aisbf_config_path = Path(args.config) / 'aisbf.json'
-        else:
-            aisbf_config_path = Path(__file__).parent / 'config' / 'aisbf.json'
-    
-    if aisbf_config_path.exists():
-        with open(aisbf_config_path) as f:
-            aisbf_config = json.load(f)
-            server_config['dashboard_config'] = aisbf_config.get('dashboard', {})
-    else:
-        server_config['dashboard_config'] = {'username': 'admin', 'password': 'admin'}
+    # Get the now-initialized server_config
+    global server_config
     
     # CLI arguments take precedence over config file
     host = args.host if args.host else server_config['host']

requirements.txt

@@ -17,3 +17,4 @@ jinja2
 itsdangerous
 bs4
 protobuf>=3.20,<4
+markdown
\ No newline at end of file

setup.py

@@ -108,6 +108,11 @@ setup(
             'templates/dashboard/index.html',
             'templates/dashboard/edit_config.html',
             'templates/dashboard/settings.html',
+            'templates/dashboard/providers.html',
+            'templates/dashboard/rotations.html',
+            'templates/dashboard/autoselect.html',
+            'templates/dashboard/prompts.html',
+            'templates/dashboard/docs.html',
         ]),
     ],
     entry_points={

templates/base.html

@@ -22,36 +22,36 @@ along with this program.  If not, see <https://www.gnu.org/licenses/>.
     <title>{% block title %}AISBF Dashboard{% endblock %}</title>
     <style>
         * { margin: 0; padding: 0; box-sizing: border-box; }
-        body { font-family: -apple-system, BlinkMacSystemFont, 'Segoe UI', Roboto, Oxygen, Ubuntu, Cantarell, sans-serif; background: #f5f5f5; }
+        body { font-family: -apple-system, BlinkMacSystemFont, 'Segoe UI', Roboto, Oxygen, Ubuntu, Cantarell, sans-serif; background: #1a1a2e; color: #e0e0e0; }
         .container { max-width: 1200px; margin: 0 auto; padding: 20px; }
-        .header { background: #2c3e50; color: white; padding: 20px 0; margin-bottom: 30px; }
+        .header { background: #16213e; color: white; padding: 20px 0; margin-bottom: 30px; border-bottom: 2px solid #0f3460; }
         .header h1 { font-size: 24px; font-weight: 600; display: inline-block; }
         .header-actions { float: right; }
-        .nav { background: white; padding: 15px; border-radius: 8px; margin-bottom: 20px; box-shadow: 0 2px 4px rgba(0,0,0,0.1); }
-        .nav a { color: #2c3e50; text-decoration: none; margin-right: 20px; padding: 8px 12px; border-radius: 4px; }
-        .nav a:hover { background: #ecf0f1; }
-        .nav a.active { background: #3498db; color: white; }
-        .content { background: white; padding: 30px; border-radius: 8px; box-shadow: 0 2px 4px rgba(0,0,0,0.1); }
+        .nav { background: #16213e; padding: 15px; border-radius: 8px; margin-bottom: 20px; box-shadow: 0 2px 4px rgba(0,0,0,0.3); }
+        .nav a { color: #a0a0a0; text-decoration: none; margin-right: 20px; padding: 8px 12px; border-radius: 4px; }
+        .nav a:hover { background: #0f3460; color: #e0e0e0; }
+        .nav a.active { background: #e94560; color: white; }
+        .content { background: #16213e; padding: 30px; border-radius: 8px; box-shadow: 0 2px 4px rgba(0,0,0,0.3); }
         .form-group { margin-bottom: 20px; }
-        .form-group label { display: block; margin-bottom: 5px; font-weight: 500; color: #2c3e50; }
-        .form-group input, .form-group textarea, .form-group select { width: 100%; padding: 10px; border: 1px solid #ddd; border-radius: 4px; font-size: 14px; }
+        .form-group label { display: block; margin-bottom: 5px; font-weight: 500; color: #e0e0e0; }
+        .form-group input, .form-group textarea, .form-group select { width: 100%; padding: 10px; border: 1px solid #0f3460; border-radius: 4px; font-size: 14px; background: #1a1a2e; color: #e0e0e0; }
         .form-group textarea { min-height: 200px; font-family: 'Courier New', monospace; }
-        .btn { padding: 10px 20px; background: #3498db; color: white; border: none; border-radius: 4px; cursor: pointer; font-size: 14px; text-decoration: none; display: inline-block; margin-left: 10px; }
-        .btn:hover { background: #2980b9; }
-        .btn-secondary { background: #95a5a6; }
-        .btn-secondary:hover { background: #7f8c8d; }
+        .btn { padding: 10px 20px; background: #e94560; color: white; border: none; border-radius: 4px; cursor: pointer; font-size: 14px; text-decoration: none; display: inline-block; margin-left: 10px; }
+        .btn:hover { background: #c73e54; }
+        .btn-secondary { background: #0f3460; }
+        .btn-secondary:hover { background: #1a4a7a; }
         .btn-danger { background: #e74c3c; }
         .btn-danger:hover { background: #c0392b; }
         .btn-warning { background: #f39c12; }
         .btn-warning:hover { background: #e67e22; }
         .alert { padding: 15px; border-radius: 4px; margin-bottom: 20px; }
-        .alert-success { background: #d4edda; color: #155724; border: 1px solid #c3e6cb; }
-        .alert-error { background: #f8d7da; color: #721c24; border: 1px solid #f5c6cb; }
-        .alert-info { background: #d1ecf1; color: #0c5460; border: 1px solid #bee5eb; }
+        .alert-success { background: #1a3a2e; color: #4ade80; border: 1px solid #22c55e; }
+        .alert-error { background: #3a1a1a; color: #f87171; border: 1px solid #ef4444; }
+        .alert-info { background: #1a2a3a; color: #60a5fa; border: 1px solid #3b82f6; }
         table { width: 100%; border-collapse: collapse; margin-top: 20px; }
-        th, td { padding: 12px; text-align: left; border-bottom: 1px solid #ddd; }
-        th { background: #f8f9fa; font-weight: 600; }
-        .code { background: #f8f9fa; padding: 15px; border-radius: 4px; font-family: 'Courier New', monospace; font-size: 13px; overflow-x: auto; }
+        th, td { padding: 12px; text-align: left; border-bottom: 1px solid #0f3460; color: #e0e0e0; }
+        th { background: #0f3460; font-weight: 600; color: #e0e0e0; }
+        .code { background: #0f3460; padding: 15px; border-radius: 4px; font-family: 'Courier New', monospace; font-size: 13px; overflow-x: auto; color: #e0e0e0; }
     </style>
     <script>
         /*
@@ -72,7 +72,7 @@ along with this program.  If not, see <https://www.gnu.org/licenses/>.
          */
         function restartServer() {
             if (confirm('Are you sure you want to restart the server? This will disconnect all active connections.')) {
-                fetch('/dashboard/restart', {
+                fetch('{{ url_for(request, "/dashboard/restart") }}', {
                     method: 'POST',
                     headers: {'Content-Type': 'application/json'}
                 })
@@ -95,7 +95,7 @@ along with this program.  If not, see <https://www.gnu.org/licenses/>.
             {% if session.logged_in %}
             <div class="header-actions">
                 <button onclick="restartServer()" class="btn btn-warning">Restart Server</button>
-                <a href="/dashboard/logout" class="btn btn-secondary">Logout</a>
+                <a href="{{ url_for(request, '/dashboard/logout') }}" class="btn btn-secondary">Logout</a>
             </div>
             {% endif %}
         </div>
@@ -104,12 +104,15 @@ along with this program.  If not, see <https://www.gnu.org/licenses/>.
     {% if session.logged_in %}
     <div class="container">
         <div class="nav">
-            <a href="/dashboard" {% if request.path == '/dashboard' %}class="active"{% endif %}>Overview</a>
-            <a href="/dashboard/providers" {% if '/providers' in request.path %}class="active"{% endif %}>Providers</a>
-            <a href="/dashboard/rotations" {% if '/rotations' in request.path %}class="active"{% endif %}>Rotations</a>
-            <a href="/dashboard/autoselect" {% if '/autoselect' in request.path %}class="active"{% endif %}>Autoselect</a>
-            <a href="/dashboard/condensation" {% if '/condensation' in request.path %}class="active"{% endif %}>Condensation</a>
-            <a href="/dashboard/settings" {% if '/settings' in request.path %}class="active"{% endif %}>Settings</a>
+            <a href="{{ url_for(request, '/dashboard') }}" {% if request.path == '/dashboard' %}class="active"{% endif %}>Overview</a>
+            <a href="{{ url_for(request, '/dashboard/providers') }}" {% if '/providers' in request.path %}class="active"{% endif %}>Providers</a>
+            <a href="{{ url_for(request, '/dashboard/rotations') }}" {% if '/rotations' in request.path %}class="active"{% endif %}>Rotations</a>
+            <a href="{{ url_for(request, '/dashboard/autoselect') }}" {% if '/autoselect' in request.path %}class="active"{% endif %}>Autoselect</a>
+            <a href="{{ url_for(request, '/dashboard/prompts') }}" {% if '/prompts' in request.path %}class="active"{% endif %}>Prompts</a>
+            <a href="{{ url_for(request, '/dashboard/settings') }}" {% if '/settings' in request.path %}class="active"{% endif %}>Settings</a>
+            <a href="{{ url_for(request, '/dashboard/docs') }}" {% if '/docs' in request.path %}class="active"{% endif %}>Docs</a>
+            <a href="{{ url_for(request, '/dashboard/about') }}" {% if '/about' in request.path %}class="active"{% endif %}>About</a>
+            <a href="{{ url_for(request, '/dashboard/license') }}" {% if '/license' in request.path %}class="active"{% endif %}>License</a>
         </div>
     </div>
     {% endif %}

templates/dashboard/autoselect.html

+<!--
+Copyright (C) 2026 Stefy Lanza <stefy@nexlab.net>
+
+This program is free software: you can redistribute it and/or modify
+it under the terms of the GNU General Public License as published by
+the Free Software Foundation, either version 3 of the License, or
+(at your option) any later version.
+
+This program is distributed in the hope that it will be useful,
+but WITHOUT ANY WARRANTY; without even the implied warranty of
+MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+GNU General Public License for more details.
+
+You should have received a copy of the GNU General Public License
+along with this program.  If not, see <https://www.gnu.org/licenses/>.
+-->
+{% extends "base.html" %}
+
+{% block title %}Autoselect - AISBF Dashboard{% endblock %}
+
+{% block content %}
+<h2 style="margin-bottom: 30px;">Autoselect Configuration</h2>
+
+{% if success %}
+<div class="alert alert-success">{{ success }}</div>
+{% endif %}
+
+{% if error %}
+<div class="alert alert-error">{{ error }}</div>
+{% endif %}
+
+<div id="autoselect-list" style="margin-bottom: 20px;">
+    <!-- Autoselect list will be rendered here -->
+</div>
+
+<button type="button" class="btn" onclick="addAutoselect()" style="margin-top: 20px;">Add Autoselect</button>
+<div style="display: flex; gap: 10px; margin-top: 20px;">
+    <button type="button" class="btn" onclick="saveAutoselect()">Save Configuration</button>
+    <a href="/dashboard" class="btn btn-secondary">Cancel</a>
+</div>
+
+<script>
+let autoselectConfig = {{ autoselect_json | safe }};
+let availableRotations = {{ available_rotations | safe }};
+let availableModels = {{ available_models | safe }};
+let expandedAutoselects = new Set();
+
+function renderAutoselectList() {
+    const container = document.getElementById('autoselect-list');
+    container.innerHTML = '';
+    
+    if (!autoselectConfig || Object.keys(autoselectConfig).length === 0) {
+        container.innerHTML = '<p style="color: #a0a0a0;">No autoselect configurations defined</p>';
+        return;
+    }
+    
+    Object.entries(autoselectConfig).forEach(([key, autoselect]) => {
+        const autoselectItem = document.createElement('div');
+        autoselectItem.className = 'autoselect-item';
+        autoselectItem.style.cssText = 'border: 1px solid #0f3460; margin-bottom: 10px; border-radius: 5px; background: #1a1a2e;';
+        
+        const isExpanded = expandedAutoselects.has(key);
+        const modelCount = autoselect.available_models ? autoselect.available_models.length : 0;
+        
+        autoselectItem.innerHTML = `
+            <div class="autoselect-header" onclick="toggleAutoselect('${key}')" style="padding: 15px; cursor: pointer; display: flex; justify-content: space-between; align-items: center; user-select: none;">
+                <div style="display: flex; align-items: center; gap: 10px;">
+                    <span style="font-size: 18px;">${isExpanded ? '▼' : '▶'}</span>
+                    <strong style="font-size: 16px;">${autoselect.model_name || key}</strong>
+                    <span style="color: #a0a0a0; font-size: 14px;">(${modelCount} available model${modelCount !== 1 ? 's' : ''})</span>
+                </div>
+                <button type="button" class="btn btn-secondary" onclick="event.stopPropagation(); removeAutoselect('${key}')" style="background: #dc3545; padding: 5px 15px;">Remove</button>
+            </div>
+            <div id="autoselect-details-${key}" style="display: ${isExpanded ? 'block' : 'none'}; padding: 20px; border-top: 1px solid #0f3460; background: #16213e;">
+                <!-- Details will be rendered here -->
+            </div>
+        `;
+        
+        container.appendChild(autoselectItem);
+        
+        if (isExpanded) {
+            renderAutoselectDetails(key);
+        }
+    });
+}
+
+function toggleAutoselect(key) {
+    if (expandedAutoselects.has(key)) {
+        expandedAutoselects.delete(key);
+    } else {
+        expandedAutoselects.add(key);
+    }
+    renderAutoselectList();
+}
+
+function renderAutoselectDetails(autoselectKey) {
+    const container = document.getElementById(`autoselect-details-${autoselectKey}`);
+    const autoselect = autoselectConfig[autoselectKey];
+    
+    // Default to "internal" if selection_model is not set
+    const selectionValue = autoselect.selection_model || 'internal';
+    
+    // Build selection model options: internal + rotations + provider models
+    let selectionOptions = `<option value="internal" ${selectionValue === 'internal' ? 'selected' : ''}>internal (Use configured internal model)</option>`;
+    selectionOptions += '<optgroup label="Rotations">';
+    selectionOptions += availableRotations.map(r =>
+        `<option value="${r}" ${selectionValue === r ? 'selected' : ''}>${r}</option>`
+    ).join('');
+    selectionOptions += '</optgroup>';
+    selectionOptions += '<optgroup label="Provider Models">';
+    selectionOptions += availableModels.map(m =>
+        `<option value="${m.id}" ${selectionValue === m.id ? 'selected' : ''}>${m.name}</option>`
+    ).join('');
+    selectionOptions += '</optgroup>';
+    
+    // Build fallback options: rotations + provider models
+    let fallbackOptions = '<optgroup label="Rotations">';
+    fallbackOptions += availableRotations.map(r =>
+        `<option value="${r}" ${autoselect.fallback === r ? 'selected' : ''}>${r}</option>`
+    ).join('');
+    fallbackOptions += '</optgroup>';
+    fallbackOptions += '<optgroup label="Provider Models">';
+    fallbackOptions += availableModels.map(m =>
+        `<option value="${m.id}" ${autoselect.fallback === m.id ? 'selected' : ''}>${m.name}</option>`
+    ).join('');
+    fallbackOptions += '</optgroup>';
+    
+    container.innerHTML = `
+        <div class="form-group">
+            <label>Model Name</label>
+            <input type="text" value="${autoselect.model_name}" onchange="updateAutoselect('${autoselectKey}', 'model_name', this.value)" required>
+        </div>
+        
+        <div class="form-group">
+            <label>Description</label>
+            <textarea onchange="updateAutoselect('${autoselectKey}', 'description', this.value)" style="min-height: 60px;">${autoselect.description || ''}</textarea>
+        </div>
+        
+        <div class="form-group">
+            <label>Selection Model (Model to use for analysis)</label>
+            <select onchange="updateAutoselect('${autoselectKey}', 'selection_model', this.value)" required>
+                <option value="">Select model...</option>
+                ${selectionOptions}
+            </select>
+            <small style="color: #a0a0a0; font-size: 12px; display: block; margin-top: 5px;">Choose "internal" to use the configured internal model, or select a rotation/provider model</small>
+        </div>
+        
+        <div class="form-group">
+            <label>Fallback Model (Default if selection fails)</label>
+            <select onchange="updateAutoselect('${autoselectKey}', 'fallback', this.value)" required>
+                <option value="">Select model...</option>
+                ${fallbackOptions}
+            </select>
+            <small style="color: #a0a0a0; font-size: 12px; display: block; margin-top: 5px;">Choose from rotations or provider models</small>
+        </div>
+        
+        <h4 style="margin-top: 20px; margin-bottom: 10px;">Available Models</h4>
+        <p style="color: #a0a0a0; font-size: 14px; margin-bottom: 10px;">Define which models can be selected and their descriptions for AI analysis</p>
+        <div id="models-${autoselectKey}"></div>
+        <button type="button" class="btn btn-secondary" onclick="addAutoselectModel('${autoselectKey}')" style="margin-top: 10px;">Add Model</button>
+    `;
+    
+    renderAutoselectModels(autoselectKey);
+}
+
+function renderAutoselectModels(autoselectKey) {
+    const container = document.getElementById(`models-${autoselectKey}`);
+    const autoselect = autoselectConfig[autoselectKey];
+    
+    if (!autoselect.available_models || autoselect.available_models.length === 0) {
+        container.innerHTML = '<p style="color: #a0a0a0;">No models configured</p>';
+        return;
+    }
+    
+    container.innerHTML = '';
+    autoselect.available_models.forEach((model, index) => {
+        const modelDiv = document.createElement('div');
+        modelDiv.style.cssText = 'border: 1px solid #0f3460; padding: 15px; margin-bottom: 10px; border-radius: 3px; background: #1a1a2e;';
+        
+        const modelOptions = availableModels.map(m =>
+            `<option value="${m.id}" ${model.model_id === m.id ? 'selected' : ''}>${m.name}</option>`
+        ).join('');
+        
+        modelDiv.innerHTML = `
+            <div style="display: flex; justify-content: space-between; align-items: center; margin-bottom: 10px;">
+                <strong>Model ${index + 1}</strong>
+                <button type="button" class="btn btn-secondary" onclick="removeAutoselectModel('${autoselectKey}', ${index})" style="background: #dc3545; padding: 5px 10px; font-size: 12px;">Remove</button>
+            </div>
+            
+            <div class="form-group">
+                <label>Model ID (Rotation or Provider Model)</label>
+                <select onchange="updateAutoselectModel('${autoselectKey}', ${index}, 'model_id', this.value)" required>
+                    <option value="">Select model...</option>
+                    ${modelOptions}
+                </select>
+                <small style="color: #a0a0a0; font-size: 12px; display: block; margin-top: 5px;">Choose from rotations or provider models</small>
+            </div>
+            
+            <div class="form-group">
+                <label>Description (Used by AI to select appropriate model)</label>
+                <textarea onchange="updateAutoselectModel('${autoselectKey}', ${index}, 'description', this.value)" style="min-height: 80px;" required>${model.description || ''}</textarea>
+                <small style="color: #666; font-size: 12px;">Be specific about when this model should be used (e.g., "Best for programming, code generation, debugging")</small>
+            </div>
+        `;
+        
+        container.appendChild(modelDiv);
+    });
+}
+
+function addAutoselect() {
+    const key = prompt('Enter autoselect key (e.g., "autoselect", "smart-select"):');
+    if (!key) {
+        return;
+    }
+    
+    if (autoselectConfig[key]) {
+        alert('Autoselect key already exists');
+        return;
+    }
+    
+    autoselectConfig[key] = {
+        model_name: key,
+        description: '',
+        selection_model: '',
+        fallback: '',
+        available_models: []
+    };
+    
+    expandedAutoselects.add(key);
+    renderAutoselectList();
+}
+
+function removeAutoselect(key) {
+    if (confirm(`Remove autoselect "${key}"?`)) {
+        delete autoselectConfig[key];
+        expandedAutoselects.delete(key);
+        renderAutoselectList();
+    }
+}
+
+function updateAutoselect(key, field, value) {
+    autoselectConfig[key][field] = value;
+}
+
+function addAutoselectModel(autoselectKey) {
+    if (!autoselectConfig[autoselectKey].available_models) {
+        autoselectConfig[autoselectKey].available_models = [];
+    }
+    
+    autoselectConfig[autoselectKey].available_models.push({
+        model_id: '',
+        description: ''
+    });
+    
+    renderAutoselectModels(autoselectKey);
+}
+
+function removeAutoselectModel(autoselectKey, index) {
+    if (confirm('Remove this model?')) {
+        autoselectConfig[autoselectKey].available_models.splice(index, 1);
+        renderAutoselectModels(autoselectKey);
+    }
+}
+
+function updateAutoselectModel(autoselectKey, index, field, value) {
+    autoselectConfig[autoselectKey].available_models[index][field] = value;
+}
+
+async function saveAutoselect() {
+    try {
+        const response = await fetch('/dashboard/autoselect', {
+            method: 'POST',
+            headers: {
+                'Content-Type': 'application/x-www-form-urlencoded',
+            },
+            body: 'config=' + encodeURIComponent(JSON.stringify(autoselectConfig, null, 2))
+        });
+        
+        if (response.ok) {
+            window.location.href = '/dashboard/autoselect?success=1';
+        } else {
+            alert('Error saving configuration');
+        }
+    } catch (error) {
+        alert('Error: ' + error.message);
+    }
+}
+
+// Initial render
+renderAutoselectList();
+</script>
+
+<style>
+.autoselect-item {
+    animation: fadeIn 0.3s;
+}
+
+.autoselect-header:hover {
+    background: #0f3460;
+}
+
+@keyframes fadeIn {
+    from { opacity: 0; transform: translateY(-10px); }
+    to { opacity: 1; transform: translateY(0); }
+}
+
+.form-group {
+    margin-bottom: 15px;
+}
+
+.form-group label {
+    display: block;
+    margin-bottom: 5px;
+    font-weight: 500;
+}
+
+.form-group input[type="text"],
+.form-group textarea,
+.form-group select {
+    width: 100%;
+    padding: 8px;
+    border: 1px solid #ddd;
+    border-radius: 3px;
+    font-size: 14px;
+}
+
+.form-group input[type="checkbox"] {
+    margin-right: 5px;
+}
+</style>
+{% endblock %}

templates/dashboard/docs.html

+<!--
+Copyright (C) 2026 Stefy Lanza <stefy@nexlab.net>
+
+This program is free software: you can redistribute it and/or modify
+it under the terms of the GNU General Public License as published by
+the Free Software Foundation, either version 3 of the License, or
+(at your option) any later version.
+
+This program is distributed in the hope that it will be useful,
+but WITHOUT ANY WARRANTY; without even the implied warranty of
+MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+GNU General Public License for more details.
+
+You should have received a copy of the GNU General Public License
+along with this program.  If not, see <https://www.gnu.org/licenses/>.
+-->
+{% extends "base.html" %}
+
+{% block title %}{{ title }} - AISBF Dashboard{% endblock %}
+
+{% block content %}
+<style>
+    .markdown-content {
+        font-family: -apple-system, BlinkMacSystemFont, 'Segoe UI', Roboto, Oxygen, Ubuntu, Cantarell, sans-serif;
+        line-height: 1.6;
+        color: #e0e0e0;
+        max-width: 100%;
+        overflow-x: auto;
+    }
+    .markdown-content pre {
+        background: #0f3460;
+        padding: 16px;
+        border-radius: 6px;
+        overflow-x: auto;
+        border: 1px solid #1a4a7a;
+        font-family: 'Courier New', Consolas, Monaco, monospace;
+        font-size: 13px;
+        line-height: 1.45;
+        color: #e0e0e0;
+    }
+    .markdown-content code {
+        background: #0f3460;
+        padding: 2px 6px;
+        border-radius: 3px;
+        font-family: 'Courier New', Consolas, Monaco, monospace;
+        font-size: 13px;
+        color: #e0e0e0;
+    }
+    .markdown-content h1 {
+        font-size: 2em;
+        border-bottom: 2px solid #e94560;
+        padding-bottom: 0.3em;
+        margin-top: 24px;
+        margin-bottom: 16px;
+        color: #ffffff;
+    }
+    .markdown-content h2 {
+        font-size: 1.5em;
+        border-bottom: 1px solid #0f3460;
+        padding-bottom: 0.3em;
+        margin-top: 24px;
+        margin-bottom: 16px;
+        color: #ffffff;
+    }
+    .markdown-content h3 {
+        font-size: 1.25em;
+        margin-top: 24px;
+        margin-bottom: 16px;
+        color: #ffffff;
+    }
+    .markdown-content h4, .markdown-content h5, .markdown-content h6 {
+        color: #ffffff;
+    }
+    .markdown-content ul, .markdown-content ol {
+        padding-left: 2em;
+        margin-bottom: 16px;
+    }
+    .markdown-content li {
+        margin-bottom: 4px;
+    }
+    .markdown-content p {
+        margin-bottom: 16px;
+    }
+    .markdown-content a {
+        color: #e94560;
+        text-decoration: none;
+    }
+    .markdown-content a:hover {
+        text-decoration: underline;
+    }
+    .markdown-content blockquote {
+        padding: 0 1em;
+        color: #a0a0a0;
+        border-left: 4px solid #e94560;
+        margin-bottom: 16px;
+    }
+    .markdown-content table {
+        border-collapse: collapse;
+        width: 100%;
+        margin-bottom: 16px;
+    }
+    .markdown-content table th,
+    .markdown-content table td {
+        padding: 6px 13px;
+        border: 1px solid #0f3460;
+    }
+    .markdown-content table th {
+        background: #0f3460;
+        font-weight: 600;
+    }
+    .markdown-content table tr:nth-child(2n) {
+        background: #0f346055;
+    }
+    .markdown-content hr {
+        border: none;
+        border-top: 1px solid #0f3460;
+        margin: 24px 0;
+    }
+    .markdown-content img {
+        max-width: 100%;
+        border-radius: 4px;
+    }
+    /* Inline code that isn't in a pre block */
+    .markdown-content p > code,
+    .markdown-content li > code {
+        background: #0f3460;
+        padding: 2px 6px;
+        border-radius: 3px;
+        color: #e94560;
+    }
+</style>
+
+<h2>{{ title }}</h2>
+
+<div class="markdown-content">
+    {{ content|safe }}
+</div>
+{% endblock %}

templates/dashboard/index.html

@@ -61,9 +61,11 @@ along with this program.  If not, see <https://www.gnu.org/licenses/>.
 </table>
 
 <h3 style="margin-top: 30px; margin-bottom: 15px;">Quick Actions</h3>
-<div style="display: flex; gap: 10px;">
+<div style="display: flex; gap: 10px; flex-wrap: wrap;">
     <a href="/dashboard/providers" class="btn">Manage Providers</a>
     <a href="/dashboard/rotations" class="btn">Manage Rotations</a>
+    <a href="/dashboard/autoselect" class="btn">Manage Autoselect</a>
+    <a href="/dashboard/prompts" class="btn">Manage Prompts</a>
     <a href="/dashboard/settings" class="btn btn-secondary">Server Settings</a>
 </div>
 {% endblock %}

templates/dashboard/prompts.html

+<!--
+Copyright (C) 2026 Stefy Lanza <stefy@nexlab.net>
+
+This program is free software: you can redistribute it and/or modify
+it under the terms of the GNU General Public License as published by
+the Free Software Foundation, either version 3 of the License, or
+(at your option) any later version.
+
+This program is distributed in the hope that it will be useful,
+but WITHOUT ANY WARRANTY; without even the implied warranty of
+MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+GNU General Public License for more details.
+
+You should have received a copy of the GNU General Public License
+along with this program.  If not, see <https://www.gnu.org/licenses/>.
+-->
+{% extends "base.html" %}
+
+{% block title %}Prompts - AISBF Dashboard{% endblock %}
+
+{% block content %}
+<h2 style="margin-bottom: 30px;">Prompts Configuration</h2>
+
+{% if success %}
+<div class="alert alert-success">{{ success }}</div>
+{% endif %}
+
+{% if error %}
+<div class="alert alert-error">{{ error }}</div>
+{% endif %}
+
+<div style="margin-bottom: 20px;">
+    <label style="font-weight: 500; margin-bottom: 10px; display: block;">Select Prompt File:</label>
+    <select id="prompt-selector" onchange="switchPrompt()" style="width: 100%; padding: 10px; border: 1px solid #ddd; border-radius: 3px; font-size: 14px;">
+        {% for prompt in prompts %}
+        <option value="{{ prompt.key }}" {% if loop.first %}selected{% endif %}>{{ prompt.name }}</option>
+        {% endfor %}
+    </select>
+</div>
+
+<form method="POST" id="prompt-form">
+    <input type="hidden" name="prompt_key" id="prompt_key" value="">
+    
+    <div class="form-group">
+        <label for="prompt_content" style="font-weight: 500; margin-bottom: 10px; display: block;">Prompt Content:</label>
+        <textarea id="prompt_content" name="prompt_content" style="width: 100%; min-height: 400px; padding: 10px; border: 1px solid #ddd; border-radius: 3px; font-family: monospace; font-size: 13px; line-height: 1.5;"></textarea>
+        <small style="color: #666; display: block; margin-top: 5px;">Edit the prompt template. Use markdown formatting as needed.</small>
+    </div>
+    
+    <div style="display: flex; gap: 10px; margin-top: 20px;">
+        <button type="submit" class="btn">Save Prompt</button>
+        <a href="/dashboard" class="btn btn-secondary">Cancel</a>
+    </div>
+</form>
+
+<script>
+const prompts = {{ prompts_data | safe }};
+let currentPromptKey = '';
+
+function switchPrompt() {
+    const selector = document.getElementById('prompt-selector');
+    const key = selector.value;
+    currentPromptKey = key;
+    
+    const prompt = prompts.find(p => p.key === key);
+    if (prompt) {
+        document.getElementById('prompt_content').value = prompt.content;
+        document.getElementById('prompt_key').value = key;
+    }
+}
+
+// Initialize with first prompt
+if (prompts.length > 0) {
+    currentPromptKey = prompts[0].key;
+    document.getElementById('prompt_content').value = prompts[0].content;
+    document.getElementById('prompt_key').value = prompts[0].key;
+}
+</script>
+
+<style>
+.form-group {
+    margin-bottom: 20px;
+}
+
+textarea {
+    resize: vertical;
+}
+
+.btn {
+    padding: 10px 20px;
+    border: none;
+    border-radius: 3px;
+    cursor: pointer;
+    font-size: 14px;
+    text-decoration: none;
+    display: inline-block;
+}
+
+.btn-secondary {
+    background: #6c757d;
+    color: white;
+}
+
+.btn-secondary:hover {
+    background: #5a6268;
+}
+</style>
+{% endblock %}

templates/dashboard/providers.html

+<!--
+Copyright (C) 2026 Stefy Lanza <stefy@nexlab.net>
+
+This program is free software: you can redistribute it and/or modify
+it under the terms of the GNU General Public License as published by
+the Free Software Foundation, either version 3 of the License, or
+(at your option) any later version.
+
+This program is distributed in the hope that it will be useful,
+but WITHOUT ANY WARRANTY; without even the implied warranty of
+MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+GNU General Public License for more details.
+
+You should have received a copy of the GNU General Public License
+along with this program.  If not, see <https://www.gnu.org/licenses/>.
+-->
+{% extends "base.html" %}
+
+{% block title %}Providers - AISBF Dashboard{% endblock %}
+
+{% block content %}
+<h2 style="margin-bottom: 30px;">Providers Configuration</h2>
+
+{% if success %}
+<div class="alert alert-success">{{ success }}</div>
+{% endif %}
+
+{% if error %}
+<div class="alert alert-error">{{ error }}</div>
+{% endif %}
+
+<!-- New Provider Form (hidden by default) -->
+<div id="new-provider-form" style="display: none; background: #16213e; border: 2px solid #4a9eff; border-radius: 8px; padding: 20px; margin-bottom: 20px;">
+    <h3 style="margin: 0 0 20px 0; color: #4a9eff;">Add New Provider</h3>
+    
+    <div class="form-group">
+        <label>Provider Key (unique identifier, e.g., "gemini", "openai", "kiro")</label>
+        <input type="text" id="new-provider-key" placeholder="Enter provider key" style="width: 100%;">
+        <small style="color: #a0a0a0; display: block; margin-top: 5px;">This will be used as the provider ID in the configuration and API endpoints</small>
+    </div>
+    
+    <div class="form-group">
+        <label>Provider Type</label>
+        <select id="new-provider-type" onchange="updateNewProviderDefaults()">
+            <option value="openai">OpenAI</option>
+            <option value="google">Google</option>
+            <option value="anthropic">Anthropic</option>
+            <option value="ollama">Ollama</option>
+            <option value="kiro">Kiro (Amazon Q Developer)</option>
+        </select>
+        <small style="color: #a0a0a0; display: block; margin-top: 5px;">Select the type of provider to configure appropriate settings</small>
+    </div>
+    
+    <div id="new-provider-type-info" style="background: #0f2840; padding: 15px; border-radius: 5px; margin-bottom: 15px; border-left: 3px solid #4a9eff;">
+        <p style="margin: 0; color: #a0a0a0;" id="new-provider-type-description">
+            OpenAI-compatible provider. Uses standard API key authentication.
+        </p>
+    </div>
+    
+    <div style="display: flex; gap: 10px;">
+        <button type="button" class="btn" onclick="confirmAddProvider()">Create Provider</button>
+        <button type="button" class="btn btn-secondary" onclick="cancelAddProvider()">Cancel</button>
+    </div>
+</div>
+
+<div id="providers-list" style="margin-bottom: 20px;">
+    <!-- Provider list will be rendered here -->
+</div>
+
+<button type="button" class="btn" id="add-provider-btn" onclick="showAddProviderForm()" style="margin-top: 20px;">Add Provider</button>
+<div style="display: flex; gap: 10px; margin-top: 20px;">
+    <button type="button" class="btn" onclick="saveProviders()">Save Configuration</button>
+    <a href="/dashboard" class="btn btn-secondary">Cancel</a>
+</div>
+
+<script>
+let providersData = {{ providers_json | safe }};
+let expandedProviders = new Set();
+
+function renderProvidersList() {
+    const container = document.getElementById('providers-list');
+    container.innerHTML = '';
+    
+    Object.entries(providersData).forEach(([key, provider]) => {
+        const providerItem = document.createElement('div');
+        providerItem.className = 'provider-item';
+        providerItem.style.cssText = 'border: 1px solid #0f3460; margin-bottom: 10px; border-radius: 5px; background: #1a1a2e;';
+        
+        const isExpanded = expandedProviders.has(key);
+        
+        providerItem.innerHTML = `
+            <div class="provider-header" onclick="toggleProvider('${key}')" style="padding: 15px; cursor: pointer; display: flex; justify-content: space-between; align-items: center; user-select: none;">
+                <div style="display: flex; align-items: center; gap: 10px;">
+                    <span style="font-size: 18px;">${isExpanded ? '▼' : '▶'}</span>
+                    <strong style="font-size: 16px;">${key}</strong>
+                    <span style="color: #a0a0a0; font-size: 14px;">(${provider.name || key})</span>
+                </div>
+                <button type="button" class="btn btn-secondary" onclick="event.stopPropagation(); removeProvider('${key}')" style="background: #dc3545; padding: 5px 15px;">Remove</button>
+            </div>
+            <div id="provider-details-${key}" style="display: ${isExpanded ? 'block' : 'none'}; padding: 20px; border-top: 1px solid #0f3460; background: #16213e;">
+                <!-- Details will be rendered here -->
+            </div>
+        `;
+        
+        container.appendChild(providerItem);
+        
+        if (isExpanded) {
+            renderProviderDetails(key);
+        }
+    });
+}
+
+function toggleProvider(key) {
+    if (expandedProviders.has(key)) {
+        expandedProviders.delete(key);
+    } else {
+        expandedProviders.add(key);
+    }
+    renderProvidersList();
+}
+
+function renderProviderDetails(key) {
+    const container = document.getElementById(`provider-details-${key}`);
+    const provider = providersData[key];
+    const isKiroProvider = provider.type === 'kiro';
+    
+    // Initialize kiro_config if this is a kiro provider and doesn't have it
+    if (isKiroProvider && !provider.kiro_config) {
+        provider.kiro_config = {
+            region: 'us-east-1',
+            creds_file: '',
+            sqlite_db: '',
+            refresh_token: '',
+            profile_arn: '',
+            client_id: '',
+            client_secret: ''
+        };
+    }
+    
+    const kiroConfig = provider.kiro_config || {};
+    
+    // Build authentication fields based on provider type
+    let authFieldsHtml = '';
+    if (isKiroProvider) {
+        // Kiro-specific authentication fields
+        authFieldsHtml = `
+            <div style="background: #0f2840; padding: 15px; border-radius: 5px; margin-bottom: 15px; border-left: 3px solid #4a9eff;">
+                <h4 style="margin: 0 0 15px 0; color: #4a9eff;">Kiro Authentication</h4>
+                <small style="color: #a0a0a0; display: block; margin-bottom: 15px;">
+                    Choose one authentication method: Kiro IDE credentials (creds_file), kiro-cli database (sqlite_db), or direct credentials (refresh_token + client_id/secret).
+                </small>
+                
+                <div class="form-group">
+                    <label>AWS Region</label>
+                    <input type="text" value="${kiroConfig.region || 'us-east-1'}" onchange="updateKiroConfig('${key}', 'region', this.value)" placeholder="us-east-1">
+                    <small style="color: #a0a0a0; display: block; margin-top: 5px;">AWS region for Kiro API (default: us-east-1)</small>
+                </div>
+                
+                <h5 style="margin: 20px 0 10px 0; color: #8ec8ff;">Option 1: Kiro IDE Credentials</h5>
+                <div class="form-group">
+                    <label>Credentials File Path</label>
+                    <input type="text" value="${kiroConfig.creds_file || ''}" onchange="updateKiroConfig('${key}', 'creds_file', this.value)" placeholder="~/.config/Code/User/globalStorage/amazon.q/credentials.json">
+                    <small style="color: #a0a0a0; display: block; margin-top: 5px;">Path to Kiro IDE credentials JSON file</small>
+                </div>
+                
+                <h5 style="margin: 20px 0 10px 0; color: #8ec8ff;">Option 2: kiro-cli Database</h5>
+                <div class="form-group">
+                    <label>SQLite Database Path</label>
+                    <input type="text" value="${kiroConfig.sqlite_db || ''}" onchange="updateKiroConfig('${key}', 'sqlite_db', this.value)" placeholder="~/.local/share/kiro-cli/data.sqlite3">
+                    <small style="color: #a0a0a0; display: block; margin-top: 5px;">Path to kiro-cli SQLite database</small>
+                </div>
+                
+                <h5 style="margin: 20px 0 10px 0; color: #8ec8ff;">Option 3: Direct Credentials</h5>
+                <div class="form-group">
+                    <label>Refresh Token</label>
+                    <input type="password" value="${kiroConfig.refresh_token || ''}" onchange="updateKiroConfig('${key}', 'refresh_token', this.value)" placeholder="Enter refresh token">
+                    <small style="color: #a0a0a0; display: block; margin-top: 5px;">Kiro refresh token for direct authentication</small>
+                </div>
+                
+                <div class="form-group">
+                    <label>Profile ARN</label>
+                    <input type="text" value="${kiroConfig.profile_arn || ''}" onchange="updateKiroConfig('${key}', 'profile_arn', this.value)" placeholder="arn:aws:codewhisperer:us-east-1:...">
+                    <small style="color: #a0a0a0; display: block; margin-top: 5px;">AWS CodeWhisperer profile ARN (optional)</small>
+                </div>
+                
+                <div class="form-group">
+                    <label>Client ID (for AWS SSO OIDC)</label>
+                    <input type="text" value="${kiroConfig.client_id || ''}" onchange="updateKiroConfig('${key}', 'client_id', this.value)" placeholder="OAuth client ID">
+                    <small style="color: #a0a0a0; display: block; margin-top: 5px;">OAuth client ID for AWS SSO OIDC authentication</small>
+                </div>
+                
+                <div class="form-group">
+                    <label>Client Secret (for AWS SSO OIDC)</label>
+                    <input type="password" value="${kiroConfig.client_secret || ''}" onchange="updateKiroConfig('${key}', 'client_secret', this.value)" placeholder="OAuth client secret">
+                    <small style="color: #a0a0a0; display: block; margin-top: 5px;">OAuth client secret for AWS SSO OIDC authentication</small>
+                </div>
+            </div>
+        `;
+    } else {
+        // Standard API key authentication fields
+        authFieldsHtml = `
+            <div class="form-group">
+                <label>
+                    <input type="checkbox" ${provider.api_key_required ? 'checked' : ''} onchange="updateProvider('${key}', 'api_key_required', this.checked)">
+                    API Key Required
+                </label>
+            </div>
+            
+            <div class="form-group">
+                <label>API Key</label>
+                <input type="password" value="${provider.api_key || ''}" onchange="updateProvider('${key}', 'api_key', this.value)" placeholder="Enter API key">
+            </div>
+        `;
+    }
+    
+    container.innerHTML = `
+        <div class="form-group">
+            <label>Provider ID</label>
+            <input type="text" value="${provider.id}" onchange="updateProvider('${key}', 'id', this.value)" required>
+        </div>
+        
+        <div class="form-group">
+            <label>Name</label>
+            <input type="text" value="${provider.name}" onchange="updateProvider('${key}', 'name', this.value)" required>
+        </div>
+        
+        <div class="form-group">
+            <label>Endpoint</label>
+            <input type="text" value="${provider.endpoint}" onchange="updateProvider('${key}', 'endpoint', this.value)" required>
+            ${isKiroProvider ? '<small style="color: #a0a0a0; display: block; margin-top: 5px;">Typically: https://q.us-east-1.amazonaws.com</small>' : ''}
+        </div>
+        
+        <div class="form-group">
+            <label>Type</label>
+            <select onchange="updateProviderType('${key}', this.value)" required>
+                <option value="google" ${provider.type === 'google' ? 'selected' : ''}>Google</option>
+                <option value="openai" ${provider.type === 'openai' ? 'selected' : ''}>OpenAI</option>
+                <option value="anthropic" ${provider.type === 'anthropic' ? 'selected' : ''}>Anthropic</option>
+                <option value="ollama" ${provider.type === 'ollama' ? 'selected' : ''}>Ollama</option>
+                <option value="kiro" ${provider.type === 'kiro' ? 'selected' : ''}>Kiro (Amazon Q Developer)</option>
+            </select>
+        </div>
+        
+        ${authFieldsHtml}
+        
+        <div class="form-group">
+            <label>Rate Limit (seconds)</label>
+            <input type="number" value="${provider.rate_limit || 0}" onchange="updateProvider('${key}', 'rate_limit', parseFloat(this.value))" step="0.1">
+            <small style="color: #a0a0a0; display: block; margin-top: 5px;">Time delay between requests to this provider</small>
+        </div>
+        
+        <div class="form-group">
+            <label>Default Rate Limit TPM (Tokens Per Minute)</label>
+            <input type="number" value="${provider.default_rate_limit_TPM || ''}" onchange="updateProvider('${key}', 'default_rate_limit_TPM', this.value ? parseInt(this.value) : null)" placeholder="Optional">
+            <small style="color: #a0a0a0; display: block; margin-top: 5px;">Default token limit per minute for models in this provider</small>
+        </div>
+        
+        <div class="form-group">
+            <label>Default Rate Limit TPH (Tokens Per Hour)</label>
+            <input type="number" value="${provider.default_rate_limit_TPH || ''}" onchange="updateProvider('${key}', 'default_rate_limit_TPH', this.value ? parseInt(this.value) : null)" placeholder="Optional">
+            <small style="color: #a0a0a0; display: block; margin-top: 5px;">Default token limit per hour for models in this provider</small>
+        </div>
+        
+        <div class="form-group">
+            <label>Default Rate Limit TPD (Tokens Per Day)</label>
+            <input type="number" value="${provider.default_rate_limit_TPD || ''}" onchange="updateProvider('${key}', 'default_rate_limit_TPD', this.value ? parseInt(this.value) : null)" placeholder="Optional">
+            <small style="color: #a0a0a0; display: block; margin-top: 5px;">Default token limit per day for models in this provider</small>
+        </div>
+        
+        <div class="form-group">
+            <label>Default Max Request Tokens</label>
+            <input type="number" value="${provider.default_max_request_tokens || ''}" onchange="updateProvider('${key}', 'default_max_request_tokens', this.value ? parseInt(this.value) : null)" placeholder="Optional">
+        </div>
+        
+        <div class="form-group">
+            <label>Default Context Size</label>
+            <input type="number" value="${provider.default_context_size || ''}" onchange="updateProvider('${key}', 'default_context_size', this.value ? parseInt(this.value) : null)" placeholder="Optional">
+        </div>
+        
+        <div class="form-group">
+            <label>Default Condense Context (%)</label>
+            <input type="number" value="${provider.default_condense_context || ''}" onchange="updateProvider('${key}', 'default_condense_context', this.value ? parseInt(this.value) : null)" placeholder="Optional (e.g., 80)">
+        </div>
+        
+        <div class="form-group">
+            <label>Default Condense Method (conversational, semantic, hierarchical, algorithmic)</label>
+            <input type="text" value="${Array.isArray(provider.default_condense_method) ? provider.default_condense_method.join(', ') : (provider.default_condense_method || '')}" onchange="updateProviderCondenseMethod('${key}', this.value)" placeholder="e.g., semantic, conversational">
+        </div>
+        
+        <h4 style="margin-top: 20px; margin-bottom: 10px;">Models</h4>
+        <div id="models-${key}"></div>
+        <button type="button" class="btn btn-secondary" onclick="addModel('${key}')" style="margin-top: 10px;">Add Model</button>
+    `;
+    
+    renderModels(key);
+}
+
+function renderModels(providerKey) {
+    const container = document.getElementById(`models-${providerKey}`);
+    const provider = providersData[providerKey];
+    
+    if (!provider.models || provider.models.length === 0) {
+        container.innerHTML = '<p style="color: #a0a0a0;">No models configured</p>';
+        return;
+    }
+    
+    container.innerHTML = '';
+    provider.models.forEach((model, index) => {
+        const modelDiv = document.createElement('div');
+        modelDiv.style.cssText = 'border: 1px solid #0f3460; padding: 15px; margin-bottom: 10px; border-radius: 3px; background: #1a1a2e;';
+        
+        modelDiv.innerHTML = `
+            <div style="display: flex; justify-content: space-between; align-items: center; margin-bottom: 10px;">
+                <strong>Model ${index + 1}</strong>
+                <button type="button" class="btn btn-secondary" onclick="removeModel('${providerKey}', ${index})" style="background: #dc3545; padding: 5px 10px; font-size: 12px;">Remove</button>
+            </div>
+            
+            <div class="form-group">
+                <label>Model Name</label>
+                <input type="text" value="${model.name}" onchange="updateModel('${providerKey}', ${index}, 'name', this.value)" required>
+            </div>
+            
+            <div class="form-group">
+                <label>Rate Limit (seconds)</label>
+                <input type="number" value="${model.rate_limit || 0}" onchange="updateModel('${providerKey}', ${index}, 'rate_limit', parseFloat(this.value))" step="0.1">
+            </div>
+            
+            <div class="form-group">
+                <label>Max Request Tokens</label>
+                <input type="number" value="${model.max_request_tokens || ''}" onchange="updateModel('${providerKey}', ${index}, 'max_request_tokens', this.value ? parseInt(this.value) : null)" placeholder="Optional">
+            </div>
+            
+            <div class="form-group">
+                <label>Context Size</label>
+                <input type="number" value="${model.context_size || ''}" onchange="updateModel('${providerKey}', ${index}, 'context_size', this.value ? parseInt(this.value) : null)" placeholder="Optional">
+            </div>
+            
+            <div class="form-group">
+                <label>Rate Limit TPM (Tokens Per Minute)</label>
+                <input type="number" value="${model.rate_limit_TPM || ''}" onchange="updateModel('${providerKey}', ${index}, 'rate_limit_TPM', this.value ? parseInt(this.value) : null)" placeholder="Optional">
+            </div>
+            
+            <div class="form-group">
+                <label>Rate Limit TPH (Tokens Per Hour)</label>
+                <input type="number" value="${model.rate_limit_TPH || ''}" onchange="updateModel('${providerKey}', ${index}, 'rate_limit_TPH', this.value ? parseInt(this.value) : null)" placeholder="Optional">
+            </div>
+            
+            <div class="form-group">
+                <label>Rate Limit TPD (Tokens Per Day)</label>
+                <input type="number" value="${model.rate_limit_TPD || ''}" onchange="updateModel('${providerKey}', ${index}, 'rate_limit_TPD', this.value ? parseInt(this.value) : null)" placeholder="Optional">
+            </div>
+            
+            <div class="form-group">
+                <label>Condense Context (%)</label>
+                <input type="number" value="${model.condense_context || ''}" onchange="updateModel('${providerKey}', ${index}, 'condense_context', this.value ? parseInt(this.value) : null)" placeholder="Optional (default: 80)">
+            </div>
+            
+            <div class="form-group">
+                <label>Condense Method (conversational, semantic, hierarchical, algorithmic)</label>
+                <input type="text" value="${Array.isArray(model.condense_method) ? model.condense_method.join(', ') : (model.condense_method || '')}" onchange="updateModelCondenseMethod('${providerKey}', ${index}, this.value)" placeholder="e.g., semantic, conversational">
+            </div>
+        `;
+        
+        container.appendChild(modelDiv);
+    });
+}
+
+function showAddProviderForm() {
+    document.getElementById('new-provider-form').style.display = 'block';
+    document.getElementById('add-provider-btn').style.display = 'none';
+    document.getElementById('new-provider-key').value = '';
+    document.getElementById('new-provider-type').value = 'openai';
+    updateNewProviderDefaults();
+    document.getElementById('new-provider-key').focus();
+}
+
+function cancelAddProvider() {
+    document.getElementById('new-provider-form').style.display = 'none';
+    document.getElementById('add-provider-btn').style.display = 'inline-block';
+}
+
+function updateNewProviderDefaults() {
+    const providerType = document.getElementById('new-provider-type').value;
+    const descriptionEl = document.getElementById('new-provider-type-description');
+    
+    const descriptions = {
+        'openai': 'OpenAI-compatible provider. Uses standard API key authentication. Endpoint: https://api.openai.com/v1',
+        'google': 'Google AI provider (Gemini). Uses API key authentication. Endpoint: https://generativelanguage.googleapis.com/v1beta',
+        'anthropic': 'Anthropic provider (Claude). Uses API key authentication. Endpoint: https://api.anthropic.com/v1',
+        'ollama': 'Ollama local provider. No API key required by default. Endpoint: http://localhost:11434/api',
+        'kiro': 'Kiro (Amazon Q Developer) provider. Uses Kiro credentials (IDE, CLI, or direct tokens). Endpoint: https://q.us-east-1.amazonaws.com'
+    };
+    
+    descriptionEl.textContent = descriptions[providerType] || 'Standard provider configuration.';
+}
+
+function confirmAddProvider() {
+    const key = document.getElementById('new-provider-key').value.trim();
+    const providerType = document.getElementById('new-provider-type').value;
+    
+    // Validate key
+    if (!key) {
+        showFormError('Please enter a provider key');
+        document.getElementById('new-provider-key').focus();
+        return;
+    }
+    
+    if (providersData[key]) {
+        showFormError('A provider with this key already exists');
+        document.getElementById('new-provider-key').focus();
+        return;
+    }
+    
+    // Validate key format (alphanumeric, dashes, underscores only)
+    if (!/^[a-zA-Z0-9_-]+$/.test(key)) {
+        showFormError('Provider key can only contain letters, numbers, dashes, and underscores');
+        document.getElementById('new-provider-key').focus();
+        return;
+    }
+    
+    // Create base provider config
+    const newProvider = {
+        id: key,
+        name: key,
+        endpoint: '',
+        type: providerType,
+        api_key_required: providerType !== 'kiro' && providerType !== 'ollama',
+        rate_limit: 0,
+        default_rate_limit: 0,
+        models: []
+    };
+    
+    // Set type-specific defaults
+    if (providerType === 'kiro') {
+        newProvider.endpoint = 'https://q.us-east-1.amazonaws.com';
+        newProvider.name = key + ' (Amazon Q Developer)';
+        newProvider.kiro_config = {
+            region: 'us-east-1',
+            creds_file: '',
+            sqlite_db: '',
+            refresh_token: '',
+            profile_arn: '',
+            client_id: '',
+            client_secret: ''
+        };
+    } else if (providerType === 'openai') {
+        newProvider.endpoint = 'https://api.openai.com/v1';
+        newProvider.api_key = '';
+    } else if (providerType === 'google') {
+        newProvider.endpoint = 'https://generativelanguage.googleapis.com/v1beta';
+        newProvider.api_key = '';
+    } else if (providerType === 'anthropic') {
+        newProvider.endpoint = 'https://api.anthropic.com/v1';
+        newProvider.api_key = '';
+    } else if (providerType === 'ollama') {
+        newProvider.endpoint = 'http://localhost:11434/api';
+        newProvider.api_key_required = false;
+        newProvider.api_key = '';
+    } else {
+        newProvider.api_key = '';
+    }
+    
+    providersData[key] = newProvider;
+    expandedProviders.add(key);
+    
+    // Hide the form and show the button
+    cancelAddProvider();
+    
+    renderProvidersList();
+    
+    // Scroll to the new provider
+    setTimeout(() => {
+        const providerElement = document.querySelector(`[onclick*="toggleProvider('${key}')"]`);
+        if (providerElement) {
+            providerElement.scrollIntoView({ behavior: 'smooth', block: 'start' });
+        }
+    }, 100);
+}
+
+function showFormError(message) {
+    // Check if error element already exists
+    let errorEl = document.getElementById('new-provider-error');
+    if (!errorEl) {
+        errorEl = document.createElement('div');
+        errorEl.id = 'new-provider-error';
+        errorEl.className = 'alert alert-error';
+        errorEl.style.marginBottom = '15px';
+        const form = document.getElementById('new-provider-form');
+        form.insertBefore(errorEl, form.querySelector('.form-group'));
+    }
+    errorEl.textContent = message;
+    errorEl.style.display = 'block';
+    
+    // Auto-hide after 5 seconds
+    setTimeout(() => {
+        if (errorEl) {
+            errorEl.style.display = 'none';
+        }
+    }, 5000);
+}
+
+// Legacy function - kept for compatibility but redirects to new form
+function addProvider() {
+    showAddProviderForm();
+}
+
+function removeProvider(key) {
+    if (confirm(`Remove provider "${key}"?`)) {
+        delete providersData[key];
+        expandedProviders.delete(key);
+        renderProvidersList();
+    }
+}
+
+function updateProvider(key, field, value) {
+    providersData[key][field] = value;
+}
+
+function updateProviderType(key, newType) {
+    const oldType = providersData[key].type;
+    providersData[key].type = newType;
+    
+    // Handle transition to/from kiro type
+    if (newType === 'kiro' && oldType !== 'kiro') {
+        // Transitioning TO kiro: initialize kiro_config, set api_key_required to false
+        providersData[key].api_key_required = false;
+        providersData[key].kiro_config = {
+            region: 'us-east-1',
+            creds_file: '',
+            sqlite_db: '',
+            refresh_token: '',
+            profile_arn: '',
+            client_id: '',
+            client_secret: ''
+        };
+        // Set default endpoint for kiro
+        if (!providersData[key].endpoint || providersData[key].endpoint === '') {
+            providersData[key].endpoint = 'https://q.us-east-1.amazonaws.com';
+        }
+    } else if (newType !== 'kiro' && oldType === 'kiro') {
+        // Transitioning FROM kiro: remove kiro_config, set api_key_required to true
+        providersData[key].api_key_required = true;
+        delete providersData[key].kiro_config;
+    }
+    
+    // Re-render to show appropriate fields
+    renderProviderDetails(key);
+}
+
+function updateKiroConfig(key, field, value) {
+    if (!providersData[key].kiro_config) {
+        providersData[key].kiro_config = {};
+    }
+    providersData[key].kiro_config[field] = value;
+}
+
+function addModel(providerKey) {
+    if (!providersData[providerKey].models) {
+        providersData[providerKey].models = [];
+    }
+    
+    providersData[providerKey].models.push({
+        name: '',
+        rate_limit: 0
+    });
+    
+    renderModels(providerKey);
+}
+
+function removeModel(providerKey, index) {
+    if (confirm('Remove this model?')) {
+        providersData[providerKey].models.splice(index, 1);
+        renderModels(providerKey);
+    }
+}
+
+function updateModel(providerKey, index, field, value) {
+    providersData[providerKey].models[index][field] = value;
+}
+
+function updateModelCondenseMethod(providerKey, index, value) {
+    const trimmed = value.trim();
+    if (!trimmed) {
+        providersData[providerKey].models[index].condense_method = null;
+        return;
+    }
+    
+    // Check if it's a comma-separated list
+    if (trimmed.includes(',')) {
+        providersData[providerKey].models[index].condense_method =
+            trimmed.split(',').map(s => s.trim()).filter(s => s);
+    } else {
+        providersData[providerKey].models[index].condense_method = trimmed;
+    }
+}
+
+function updateProviderCondenseMethod(providerKey, value) {
+    const trimmed = value.trim();
+    if (!trimmed) {
+        providersData[providerKey].default_condense_method = null;
+        return;
+    }
+    
+    // Check if it's a comma-separated list
+    if (trimmed.includes(',')) {
+        providersData[providerKey].default_condense_method =
+            trimmed.split(',').map(s => s.trim()).filter(s => s);
+    } else {
+        providersData[providerKey].default_condense_method = trimmed;
+    }
+}
+
+async function saveProviders() {
+    try {
+        const response = await fetch('/dashboard/providers', {
+            method: 'POST',
+            headers: {
+                'Content-Type': 'application/x-www-form-urlencoded',
+            },
+            body: 'config=' + encodeURIComponent(JSON.stringify(providersData, null, 2))
+        });
+        
+        if (response.ok) {
+            window.location.href = '/dashboard/providers?success=1';
+        } else {
+            alert('Error saving configuration');
+        }
+    } catch (error) {
+        alert('Error: ' + error.message);
+    }
+}
+
+// Initial render
+renderProvidersList();
+</script>
+
+<style>
+.provider-item {
+    animation: fadeIn 0.3s;
+}
+
+.provider-header:hover {
+    background: #0f3460;
+}
+
+@keyframes fadeIn {
+    from { opacity: 0; transform: translateY(-10px); }
+    to { opacity: 1; transform: translateY(0); }
+}
+
+.form-group {
+    margin-bottom: 15px;
+}
+
+.form-group label {
+    display: block;
+    margin-bottom: 5px;
+    font-weight: 500;
+    color: #e0e0e0;
+}
+
+.form-group input[type="text"],
+.form-group input[type="password"],
+.form-group input[type="number"],
+.form-group select {
+    width: 100%;
+    padding: 8px;
+    border: 1px solid #0f3460;
+    border-radius: 3px;
+    font-size: 14px;
+    background: #1a1a2e;
+    color: #e0e0e0;
+}
+
+.form-group input[type="checkbox"] {
+    margin-right: 5px;
+}
+</style>
+{% endblock %}

templates/dashboard/rotations.html

+<!--
+Copyright (C) 2026 Stefy Lanza <stefy@nexlab.net>
+
+This program is free software: you can redistribute it and/or modify
+it under the terms of the GNU General Public License as published by
+the Free Software Foundation, either version 3 of the License, or
+(at your option) any later version.
+
+This program is distributed in the hope that it will be useful,
+but WITHOUT ANY WARRANTY; without even the implied warranty of
+MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+GNU General Public License for more details.
+
+You should have received a copy of the GNU General Public License
+along with this program.  If not, see <https://www.gnu.org/licenses/>.
+-->
+{% extends "base.html" %}
+
+{% block title %}Rotations - AISBF Dashboard{% endblock %}
+
+{% block content %}
+<h2 style="margin-bottom: 30px;">Rotations Configuration</h2>
+
+{% if success %}
+<div class="alert alert-success">{{ success }}</div>
+{% endif %}
+
+{% if error %}
+<div class="alert alert-error">{{ error }}</div>
+{% endif %}
+
+<div class="form-group" style="margin-bottom: 20px;">
+    <label>
+        <input type="checkbox" id="global-notify" onchange="updateGlobalNotify(this.checked)">
+        Global Notify Errors
+    </label>
+</div>
+
+<div id="rotations-list" style="margin-bottom: 20px;">
+    <!-- Rotations list will be rendered here -->
+</div>
+
+<button type="button" class="btn" onclick="addRotation()" style="margin-top: 20px;">Add Rotation</button>
+<div style="display: flex; gap: 10px; margin-top: 20px;">
+    <button type="button" class="btn" onclick="saveRotations()">Save Configuration</button>
+    <a href="/dashboard" class="btn btn-secondary">Cancel</a>
+</div>
+
+<script>
+let rotationsConfig = {{ rotations_json | safe }};
+let availableProviders = {{ available_providers | safe }};
+let expandedRotations = new Set();
+
+document.getElementById('global-notify').checked = rotationsConfig.notifyerrors || false;
+
+function renderRotationsList() {
+    const container = document.getElementById('rotations-list');
+    container.innerHTML = '';
+    
+    Object.entries(rotationsConfig.rotations || {}).forEach(([key, rotation]) => {
+        const rotationItem = document.createElement('div');
+        rotationItem.className = 'rotation-item';
+        rotationItem.style.cssText = 'border: 1px solid #0f3460; margin-bottom: 10px; border-radius: 5px; background: #1a1a2e;';
+        
+        const isExpanded = expandedRotations.has(key);
+        const providerCount = rotation.providers ? rotation.providers.length : 0;
+        
+        rotationItem.innerHTML = `
+            <div class="rotation-header" onclick="toggleRotation('${key}')" style="padding: 15px; cursor: pointer; display: flex; justify-content: space-between; align-items: center; user-select: none;">
+                <div style="display: flex; align-items: center; gap: 10px;">
+                    <span style="font-size: 18px;">${isExpanded ? '▼' : '▶'}</span>
+                    <strong style="font-size: 16px;">${key}</strong>
+                    <span style="color: #a0a0a0; font-size: 14px;">(${providerCount} provider${providerCount !== 1 ? 's' : ''})</span>
+                </div>
+                <button type="button" class="btn btn-secondary" onclick="event.stopPropagation(); removeRotation('${key}')" style="background: #dc3545; padding: 5px 15px;">Remove</button>
+            </div>
+            <div id="rotation-details-${key}" style="display: ${isExpanded ? 'block' : 'none'}; padding: 20px; border-top: 1px solid #0f3460; background: #16213e;">
+                <!-- Details will be rendered here -->
+            </div>
+        `;
+        
+        container.appendChild(rotationItem);
+        
+        if (isExpanded) {
+            renderRotationDetails(key);
+        }
+    });
+}
+
+function toggleRotation(key) {
+    if (expandedRotations.has(key)) {
+        expandedRotations.delete(key);
+    } else {
+        expandedRotations.add(key);
+    }
+    renderRotationsList();
+}
+
+function renderRotationDetails(rotationKey) {
+    const container = document.getElementById(`rotation-details-${rotationKey}`);
+    const rotation = rotationsConfig.rotations[rotationKey];
+    
+    container.innerHTML = `
+        <div class="form-group">
+            <label>Model Name</label>
+            <input type="text" value="${rotation.model_name}" onchange="updateRotation('${rotationKey}', 'model_name', this.value)" required>
+        </div>
+        
+        <div class="form-group">
+            <label>
+                <input type="checkbox" ${rotation.notifyerrors ? 'checked' : ''} onchange="updateRotation('${rotationKey}', 'notifyerrors', this.checked)">
+                Notify Errors
+            </label>
+        </div>
+        
+        <div class="form-group">
+            <label>Default Rate Limit (seconds)</label>
+            <input type="number" value="${rotation.default_rate_limit || ''}" onchange="updateRotation('${rotationKey}', 'default_rate_limit', this.value ? parseFloat(this.value) : null)" step="0.1" placeholder="Optional">
+        </div>
+        
+        <div class="form-group">
+            <label>Default Context Size</label>
+            <input type="number" value="${rotation.default_context_size || ''}" onchange="updateRotation('${rotationKey}', 'default_context_size', this.value ? parseInt(this.value) : null)" placeholder="Optional">
+        </div>
+        
+        <h4 style="margin-top: 20px; margin-bottom: 10px;">Providers</h4>
+        <div id="providers-${rotationKey}"></div>
+        <button type="button" class="btn btn-secondary" onclick="addRotationProvider('${rotationKey}')" style="margin-top: 10px;">Add Provider</button>
+    `;
+    
+    renderRotationProviders(rotationKey);
+}
+
+function renderRotationProviders(rotationKey) {
+    const container = document.getElementById(`providers-${rotationKey}`);
+    const rotation = rotationsConfig.rotations[rotationKey];
+    
+    if (!rotation.providers || rotation.providers.length === 0) {
+        container.innerHTML = '<p style="color: #a0a0a0;">No providers configured</p>';
+        return;
+    }
+    
+    container.innerHTML = '';
+    rotation.providers.forEach((provider, providerIndex) => {
+        const providerDiv = document.createElement('div');
+        providerDiv.style.cssText = 'border: 1px solid #0f3460; padding: 15px; margin-bottom: 10px; border-radius: 3px; background: #1a1a2e;';
+        
+        const providerOptions = availableProviders.map(p => 
+            `<option value="${p}" ${provider.provider_id === p ? 'selected' : ''}>${p}</option>`
+        ).join('');
+        
+        providerDiv.innerHTML = `
+            <div style="display: flex; justify-content: space-between; align-items: center; margin-bottom: 10px;">
+                <strong>Provider ${providerIndex + 1}</strong>
+                <button type="button" class="btn btn-secondary" onclick="removeRotationProvider('${rotationKey}', ${providerIndex})" style="background: #dc3545; padding: 5px 10px; font-size: 12px;">Remove</button>
+            </div>
+            
+            <div class="form-group">
+                <label>Provider ID</label>
+                <select onchange="updateRotationProvider('${rotationKey}', ${providerIndex}, 'provider_id', this.value)" required>
+                    <option value="">Select provider...</option>
+                    ${providerOptions}
+                </select>
+            </div>
+            
+            <div class="form-group">
+                <label>Weight (optional, for provider-level weight)</label>
+                <input type="number" value="${provider.weight || ''}" onchange="updateRotationProvider('${rotationKey}', ${providerIndex}, 'weight', this.value ? parseInt(this.value) : null)" placeholder="Optional">
+            </div>
+            
+            <h5 style="margin-top: 15px; margin-bottom: 10px;">Models</h5>
+            <div id="models-${rotationKey}-${providerIndex}"></div>
+            <button type="button" class="btn btn-secondary" onclick="addRotationModel('${rotationKey}', ${providerIndex})" style="margin-top: 10px; font-size: 12px;">Add Model</button>
+            <p style="font-size: 12px; color: #a0a0a0; margin-top: 5px;">Leave models empty to use all models from provider config</p>
+        `;
+        
+        container.appendChild(providerDiv);
+        renderRotationModels(rotationKey, providerIndex);
+    });
+}
+
+function renderRotationModels(rotationKey, providerIndex) {
+    const container = document.getElementById(`models-${rotationKey}-${providerIndex}`);
+    const provider = rotationsConfig.rotations[rotationKey].providers[providerIndex];
+    
+    if (!provider.models || provider.models.length === 0) {
+        container.innerHTML = '<p style="color: #a0a0a0; font-size: 12px;">No models specified (will use all from provider)</p>';
+        return;
+    }
+    
+    container.innerHTML = '';
+    provider.models.forEach((model, modelIndex) => {
+        const modelDiv = document.createElement('div');
+        modelDiv.style.cssText = 'border: 1px solid #0f3460; padding: 10px; margin-bottom: 8px; border-radius: 3px; background: #16213e;';
+        
+        modelDiv.innerHTML = `
+            <div style="display: flex; justify-content: space-between; align-items: center; margin-bottom: 8px;">
+                <strong style="font-size: 13px;">Model ${modelIndex + 1}</strong>
+                <button type="button" onclick="removeRotationModel('${rotationKey}', ${providerIndex}, ${modelIndex})" style="background: #dc3545; color: white; border: none; padding: 3px 8px; border-radius: 3px; cursor: pointer; font-size: 11px;">Remove</button>
+            </div>
+            
+            <div class="form-group" style="margin-bottom: 8px;">
+                <label style="font-size: 12px;">Model Name</label>
+                <input type="text" value="${model.name}" onchange="updateRotationModel('${rotationKey}', ${providerIndex}, ${modelIndex}, 'name', this.value)" required style="font-size: 12px; padding: 5px;">
+            </div>
+            
+            <div class="form-group" style="margin-bottom: 8px;">
+                <label style="font-size: 12px;">Weight</label>
+                <input type="number" value="${model.weight || 1}" onchange="updateRotationModel('${rotationKey}', ${providerIndex}, ${modelIndex}, 'weight', parseInt(this.value))" style="font-size: 12px; padding: 5px;">
+            </div>
+            
+            <div class="form-group" style="margin-bottom: 8px;">
+                <label style="font-size: 12px;">Rate Limit (seconds)</label>
+                <input type="number" value="${model.rate_limit || 0}" onchange="updateRotationModel('${rotationKey}', ${providerIndex}, ${modelIndex}, 'rate_limit', parseFloat(this.value))" step="0.1" style="font-size: 12px; padding: 5px;">
+            </div>
+            
+            <div class="form-group" style="margin-bottom: 8px;">
+                <label style="font-size: 12px;">Max Request Tokens</label>
+                <input type="number" value="${model.max_request_tokens || ''}" onchange="updateRotationModel('${rotationKey}', ${providerIndex}, ${modelIndex}, 'max_request_tokens', this.value ? parseInt(this.value) : null)" placeholder="Optional" style="font-size: 12px; padding: 5px;">
+            </div>
+            
+            <div class="form-group" style="margin-bottom: 8px;">
+                <label style="font-size: 12px;">Context Size</label>
+                <input type="number" value="${model.context_size || ''}" onchange="updateRotationModel('${rotationKey}', ${providerIndex}, ${modelIndex}, 'context_size', this.value ? parseInt(this.value) : null)" placeholder="Optional" style="font-size: 12px; padding: 5px;">
+            </div>
+            
+            <div class="form-group" style="margin-bottom: 8px;">
+                <label style="font-size: 12px;">Condense Context (%)</label>
+                <input type="number" value="${model.condense_context || ''}" onchange="updateRotationModel('${rotationKey}', ${providerIndex}, ${modelIndex}, 'condense_context', this.value ? parseInt(this.value) : null)" placeholder="Optional (default: 80)" style="font-size: 12px; padding: 5px;">
+            </div>
+            
+            <div class="form-group" style="margin-bottom: 0;">
+                <label style="font-size: 12px;">Condense Method</label>
+                <input type="text" value="${Array.isArray(model.condense_method) ? model.condense_method.join(', ') : (model.condense_method || '')}" onchange="updateRotationModelCondenseMethod('${rotationKey}', ${providerIndex}, ${modelIndex}, this.value)" placeholder="e.g., semantic, conversational, hierarchical" style="font-size: 12px; padding: 5px;">
+            </div>
+        `;
+        
+        container.appendChild(modelDiv);
+    });
+}
+
+function updateGlobalNotify(value) {
+    rotationsConfig.notifyerrors = value;
+}
+
+function addRotation() {
+    const key = prompt('Enter rotation key (e.g., "coding", "general"):');
+    if (!key || rotationsConfig.rotations[key]) {
+        alert('Invalid or duplicate rotation key');
+        return;
+    }
+    
+    if (!rotationsConfig.rotations) {
+        rotationsConfig.rotations = {};
+    }
+    
+    rotationsConfig.rotations[key] = {
+        model_name: key,
+        notifyerrors: false,
+        providers: []
+    };
+    
+    expandedRotations.add(key);
+    renderRotationsList();
+}
+
+function removeRotation(key) {
+    if (confirm(`Remove rotation "${key}"?`)) {
+        delete rotationsConfig.rotations[key];
+        expandedRotations.delete(key);
+        renderRotationsList();
+    }
+}
+
+function updateRotation(key, field, value) {
+    rotationsConfig.rotations[key][field] = value;
+}
+
+function addRotationProvider(rotationKey) {
+    if (!rotationsConfig.rotations[rotationKey].providers) {
+        rotationsConfig.rotations[rotationKey].providers = [];
+    }
+    
+    rotationsConfig.rotations[rotationKey].providers.push({
+        provider_id: '',
+        models: []
+    });
+    
+    renderRotationProviders(rotationKey);
+}
+
+function removeRotationProvider(rotationKey, providerIndex) {
+    if (confirm('Remove this provider?')) {
+        rotationsConfig.rotations[rotationKey].providers.splice(providerIndex, 1);
+        renderRotationProviders(rotationKey);
+    }
+}
+
+function updateRotationProvider(rotationKey, providerIndex, field, value) {
+    rotationsConfig.rotations[rotationKey].providers[providerIndex][field] = value;
+}
+
+function addRotationModel(rotationKey, providerIndex) {
+    const provider = rotationsConfig.rotations[rotationKey].providers[providerIndex];
+    if (!provider.models) {
+        provider.models = [];
+    }
+    
+    provider.models.push({
+        name: '',
+        weight: 1,
+        rate_limit: 0
+    });
+    
+    renderRotationModels(rotationKey, providerIndex);
+}
+
+function removeRotationModel(rotationKey, providerIndex, modelIndex) {
+    if (confirm('Remove this model?')) {
+        rotationsConfig.rotations[rotationKey].providers[providerIndex].models.splice(modelIndex, 1);
+        renderRotationModels(rotationKey, providerIndex);
+    }
+}
+
+function updateRotationModel(rotationKey, providerIndex, modelIndex, field, value) {
+    rotationsConfig.rotations[rotationKey].providers[providerIndex].models[modelIndex][field] = value;
+}
+
+function updateRotationModelCondenseMethod(rotationKey, providerIndex, modelIndex, value) {
+    const trimmed = value.trim();
+    if (!trimmed) {
+        rotationsConfig.rotations[rotationKey].providers[providerIndex].models[modelIndex].condense_method = null;
+        return;
+    }
+    
+    // Check if it's a comma-separated list
+    if (trimmed.includes(',')) {
+        rotationsConfig.rotations[rotationKey].providers[providerIndex].models[modelIndex].condense_method = 
+            trimmed.split(',').map(s => s.trim()).filter(s => s);
+    } else {
+        rotationsConfig.rotations[rotationKey].providers[providerIndex].models[modelIndex].condense_method = trimmed;
+    }
+}
+
+async function saveRotations() {
+    try {
+        const response = await fetch('/dashboard/rotations', {
+            method: 'POST',
+            headers: {
+                'Content-Type': 'application/x-www-form-urlencoded',
+            },
+            body: 'config=' + encodeURIComponent(JSON.stringify(rotationsConfig, null, 2))
+        });
+        
+        if (response.ok) {
+            window.location.href = '/dashboard/rotations?success=1';
+        } else {
+            alert('Error saving configuration');
+        }
+    } catch (error) {
+        alert('Error: ' + error.message);
+    }
+}
+
+// Initial render
+renderRotationsList();
+</script>
+
+<style>
+.rotation-item {
+    animation: fadeIn 0.3s;
+}
+
+.rotation-header:hover {
+    background: #0f3460;
+}
+
+@keyframes fadeIn {
+    from { opacity: 0; transform: translateY(-10px); }
+    to { opacity: 1; transform: translateY(0); }
+}
+
+.form-group {
+    margin-bottom: 15px;
+}
+
+.form-group label {
+    display: block;
+    margin-bottom: 5px;
+    font-weight: 500;
+    color: #e0e0e0;
+}
+
+.form-group input[type="text"],
+.form-group input[type="number"],
+.form-group select {
+    width: 100%;
+    padding: 8px;
+    border: 1px solid #0f3460;
+    border-radius: 3px;
+    font-size: 14px;
+    background: #1a1a2e;
+    color: #e0e0e0;
+}
+
+.form-group input[type="checkbox"] {
+    margin-right: 5px;
+}
+</style>
+{% endblock %}

templates/dashboard/settings.html

@@ -44,12 +44,32 @@ along with this program.  If not, see <https://www.gnu.org/licenses/>.
     
     <div class="form-group">
         <label for="protocol">Protocol</label>
-        <select id="protocol" name="protocol">
+        <select id="protocol" name="protocol" onchange="toggleSSLFields()">
             <option value="http" {% if config.server.protocol == 'http' %}selected{% endif %}>HTTP</option>
             <option value="https" {% if config.server.protocol == 'https' %}selected{% endif %}>HTTPS</option>
         </select>
     </div>
     
+    <div id="ssl-fields" style="display: {% if config.server.protocol == 'https' %}block{% else %}none{% endif %};">
+        <div class="form-group">
+            <label for="public_domain">Public Domain (for Let's Encrypt)</label>
+            <input type="text" id="public_domain" name="public_domain" value="{{ config.server.public_domain or '' }}" placeholder="example.com (optional)">
+            <small style="color: #666; display: block; margin-top: 5px;">If specified, Let's Encrypt will be used to generate valid SSL certificates. Leave empty to use self-signed certificates.</small>
+        </div>
+        
+        <div class="form-group">
+            <label for="ssl_cert">SSL Certificate Path</label>
+            <input type="text" id="ssl_cert" name="ssl_cert" value="{{ config.server.ssl_cert or '' }}" placeholder="~/.aisbf/cert.pem (default)">
+            <small style="color: #666; display: block; margin-top: 5px;">Path to SSL certificate file. Leave empty to use default (~/.aisbf/cert.pem). Auto-generated using Let's Encrypt if public domain is set, otherwise self-signed.</small>
+        </div>
+        
+        <div class="form-group">
+            <label for="ssl_key">SSL Key Path</label>
+            <input type="text" id="ssl_key" name="ssl_key" value="{{ config.server.ssl_key or '' }}" placeholder="~/.aisbf/key.pem (default)">
+            <small style="color: #666; display: block; margin-top: 5px;">Path to SSL private key file. Leave empty to use default (~/.aisbf/key.pem). Auto-generated with certificate.</small>
+        </div>
+    </div>
+    
     <h3 style="margin: 30px 0 20px;">Authentication</h3>
     
     <div class="form-group">
@@ -76,11 +96,18 @@ along with this program.  If not, see <https://www.gnu.org/licenses/>.
         <input type="password" id="dashboard_password" name="dashboard_password" placeholder="Leave blank to keep current">
     </div>
     
-    <h3 style="margin: 30px 0 20px;">Internal Model</h3>
+    <h3 style="margin: 30px 0 20px;">Internal Models</h3>
     
     <div class="form-group">
-        <label for="internal_model_id">Model ID</label>
-        <input type="text" id="internal_model_id" name="internal_model_id" value="{{ config.internal_model.model_id }}" required>
+        <label for="condensation_model_id">Condensation Model ID</label>
+        <input type="text" id="condensation_model_id" name="condensation_model_id" value="{{ config.internal_model.condensation_model_id }}" required>
+        <small style="color: #666; display: block; margin-top: 5px;">Used when condensation model is set to "internal"</small>
+    </div>
+    
+    <div class="form-group">
+        <label for="autoselect_model_id">Autoselect Model ID</label>
+        <input type="text" id="autoselect_model_id" name="autoselect_model_id" value="{{ config.internal_model.autoselect_model_id }}" required>
+        <small style="color: #666; display: block; margin-top: 5px;">Used when autoselect selection_model is set to "internal"</small>
     </div>
     
     <div style="display: flex; gap: 10px; margin-top: 30px;">
@@ -88,4 +115,17 @@ along with this program.  If not, see <https://www.gnu.org/licenses/>.
         <a href="/dashboard" class="btn btn-secondary">Cancel</a>
     </div>
 </form>
+
+<script>
+function toggleSSLFields() {
+    const protocol = document.getElementById('protocol').value;
+    const sslFields = document.getElementById('ssl-fields');
+    
+    if (protocol === 'https') {
+        sslFields.style.display = 'block';
+    } else {
+        sslFields.style.display = 'none';
+    }
+}
+</script>
 {% endblock %}