feat: Add intelligent 429 rate limit handling and improve configuration

- Implement intelligent 429 rate limit parsing in providers
  - Parse Retry-After and X-RateLimit-Reset headers
  - Extract wait time from response body fields
  - Parse error messages for time patterns
  - Automatically disable providers for exact duration specified

- Update aisbf.sh to read port from config file dynamically
  - Add get_port() function to parse aisbf.json
  - Remove hardcoded port 17765
  - Show port number when starting server

- Implement SHA256 password hashing for dashboard
  - Store password as hash in aisbf.json
  - Hash passwords during login validation
  - Hash passwords when updating settings

- Add templates to package distribution
  - Update setup.py to include dashboard templates
  - Update MANIFEST.in (already included templates)
  - Add aisbf.json to package

- Update documentation
  - Add OpenAI-compatible v1 endpoints to README
  - Add dashboard endpoints documentation
  - Create comprehensive API_EXAMPLES.md with examples in cURL, Python, and JavaScript

- Add GPL license headers to all template files
  - HTML templates in templates/ directory
  - JavaScript code in templates/base.html

All changes maintain backward compatibility while adding new features.

API_EXAMPLES.md

+# AISBF API Examples
+
+This document provides practical examples for using the AISBF API endpoints.
+
+## Table of Contents
+- [OpenAI-Compatible v1 Endpoints](#openai-compatible-v1-endpoints)
+- [Chat Completions](#chat-completions)
+- [Audio Endpoints](#audio-endpoints)
+- [Image Generation](#image-generation)
+- [Embeddings](#embeddings)
+- [Model Listing](#model-listing)
+- [Rotations](#rotations)
+- [Autoselect](#autoselect)
+
+## OpenAI-Compatible v1 Endpoints
+
+The v1 endpoints follow the standard OpenAI API format with `provider/model` notation.
+
+### Chat Completions
+
+#### Using cURL
+
+```bash
+curl -X POST http://localhost:17765/api/v1/chat/completions \
+  -H "Content-Type: application/json" \
+  -d '{
+    "model": "openai/gpt-4",
+    "messages": [
+      {"role": "user", "content": "Hello, how are you?"}
+    ]
+  }'
+```
+
+#### Using Python
+
+```python
+import requests
+
+response = requests.post(
+    "http://localhost:17765/api/v1/chat/completions",
+    json={
+        "model": "openai/gpt-4",
+        "messages": [
+            {"role": "user", "content": "Hello, how are you?"}
+        ]
+    }
+)
+
+print(response.json())
+```
+
+#### 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
+)
+
+response = client.chat.completions.create(
+    model="openai/gpt-4",
+    messages=[
+        {"role": "user", "content": "Hello, how are you?"}
+    ]
+)
+
+print(response.choices[0].message.content)
+```
+
+#### Streaming Response
+
+```python
+from openai import OpenAI
+
+client = OpenAI(
+    base_url="http://localhost:17765/api/v1",
+    api_key="dummy"
+)
+
+stream = client.chat.completions.create(
+    model="gemini/gemini-2.0-flash",
+    messages=[
+        {"role": "user", "content": "Write a short poem"}
+    ],
+    stream=True
+)
+
+for chunk in stream:
+    if chunk.choices[0].delta.content:
+        print(chunk.choices[0].delta.content, end="")
+```
+
+#### Using 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"}]
+  }'
+
+# 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"}]
+  }'
+```
+
+## Audio Endpoints
+
+### Audio Transcription
+
+```bash
+curl -X POST http://localhost:17765/api/v1/audio/transcriptions \
+  -F "file=@audio.mp3" \
+  -F "model=openai/whisper-1"
+```
+
+```python
+from openai import OpenAI
+
+client = OpenAI(
+    base_url="http://localhost:17765/api/v1",
+    api_key="dummy"
+)
+
+with open("audio.mp3", "rb") as audio_file:
+    transcript = client.audio.transcriptions.create(
+        model="openai/whisper-1",
+        file=audio_file
+    )
+
+print(transcript.text)
+```
+
+### Text-to-Speech
+
+```bash
+curl -X POST http://localhost:17765/api/v1/audio/speech \
+  -H "Content-Type: application/json" \
+  -d '{
+    "model": "openai/tts-1",
+    "input": "Hello, this is a test.",
+    "voice": "alloy"
+  }' \
+  --output speech.mp3
+```
+
+```python
+from openai import OpenAI
+
+client = OpenAI(
+    base_url="http://localhost:17765/api/v1",
+    api_key="dummy"
+)
+
+response = client.audio.speech.create(
+    model="openai/tts-1",
+    voice="alloy",
+    input="Hello, this is a test."
+)
+
+response.stream_to_file("speech.mp3")
+```
+
+## Image Generation
+
+```bash
+curl -X POST http://localhost:17765/api/v1/images/generations \
+  -H "Content-Type: application/json" \
+  -d '{
+    "model": "openai/dall-e-3",
+    "prompt": "A beautiful sunset over mountains",
+    "n": 1,
+    "size": "1024x1024"
+  }'
+```
+
+```python
+from openai import OpenAI
+
+client = OpenAI(
+    base_url="http://localhost:17765/api/v1",
+    api_key="dummy"
+)
+
+response = client.images.generate(
+    model="openai/dall-e-3",
+    prompt="A beautiful sunset over mountains",
+    n=1,
+    size="1024x1024"
+)
+
+print(response.data[0].url)
+```
+
+## Embeddings
+
+```bash
+curl -X POST http://localhost:17765/api/v1/embeddings \
+  -H "Content-Type: application/json" \
+  -d '{
+    "model": "openai/text-embedding-ada-002",
+    "input": "The quick brown fox jumps over the lazy dog"
+  }'
+```
+
+```python
+from openai import OpenAI
+
+client = OpenAI(
+    base_url="http://localhost:17765/api/v1",
+    api_key="dummy"
+)
+
+response = client.embeddings.create(
+    model="openai/text-embedding-ada-002",
+    input="The quick brown fox jumps over the lazy dog"
+)
+
+print(response.data[0].embedding)
+```
+
+## Model Listing
+
+### List All Models
+
+```bash
+curl http://localhost:17765/api/v1/models
+```
+
+```python
+from openai import OpenAI
+
+client = OpenAI(
+    base_url="http://localhost:17765/api/v1",
+    api_key="dummy"
+)
+
+models = client.models.list()
+for model in models.data:
+    print(f"{model.id} - {model.owned_by}")
+```
+
+## Rotations
+
+Rotations provide weighted load balancing across multiple providers.
+
+### Using Rotation with v1 Endpoint
+
+```bash
+curl -X POST http://localhost:17765/api/v1/chat/completions \
+  -H "Content-Type: application/json" \
+  -d '{
+    "model": "coding",
+    "messages": [
+      {"role": "user", "content": "Write a Python function to sort a list"}
+    ]
+  }'
+```
+
+### 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 Available Rotations
+
+```bash
+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 \
+  -H "Content-Type: application/json" \
+  -d '{
+    "model": "autoselect",
+    "messages": [
+      {"role": "user", "content": "Debug this Python code: def add(a,b): return a-b"}
+    ]
+  }'
+```
+
+### List Available Autoselect Configurations
+
+```bash
+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 \
+  -H "Content-Type: application/json" \
+  -d '{
+    "model": "gpt-4",
+    "messages": [{"role": "user", "content": "Hello"}]
+  }'
+
+# List provider models
+curl http://localhost:17765/api/openai/models
+```
+
+## Authentication
+
+If authentication is enabled in your configuration:
+
+```bash
+curl -X POST http://localhost:17765/api/v1/chat/completions \
+  -H "Content-Type: application/json" \
+  -H "Authorization: Bearer YOUR_TOKEN_HERE" \
+  -d '{
+    "model": "openai/gpt-4",
+    "messages": [{"role": "user", "content": "Hello"}]
+  }'
+```
+
+```python
+from openai import OpenAI
+
+client = OpenAI(
+    base_url="http://localhost:17765/api/v1",
+    api_key="YOUR_TOKEN_HERE"
+)
+
+response = client.chat.completions.create(
+    model="openai/gpt-4",
+    messages=[{"role": "user", "content": "Hello"}]
+)
+```
+
+## JavaScript/Node.js Examples
+
+### Using fetch API
+
+```javascript
+const response = await fetch('http://localhost:17765/api/v1/chat/completions', {
+  method: 'POST',
+  headers: {
+    'Content-Type': 'application/json',
+  },
+  body: JSON.stringify({
+    model: 'openai/gpt-4',
+    messages: [
+      { role: 'user', content: 'Hello, how are you?' }
+    ]
+  })
+});
+
+const data = await response.json();
+console.log(data.choices[0].message.content);
+```
+
+### Using OpenAI SDK
+
+```javascript
+import OpenAI from 'openai';
+
+const client = new OpenAI({
+  baseURL: 'http://localhost:17765/api/v1',
+  apiKey: 'dummy'
+});
+
+const response = await client.chat.completions.create({
+  model: 'openai/gpt-4',
+  messages: [
+    { role: 'user', content: 'Hello, how are you?' }
+  ]
+});
+
+console.log(response.choices[0].message.content);
+```
+
+### Streaming in JavaScript
+
+```javascript
+import OpenAI from 'openai';
+
+const client = new OpenAI({
+  baseURL: 'http://localhost:17765/api/v1',
+  apiKey: 'dummy'
+});
+
+const stream = await client.chat.completions.create({
+  model: 'gemini/gemini-2.0-flash',
+  messages: [
+    { role: 'user', content: 'Write a short poem' }
+  ],
+  stream: true
+});
+
+for await (const chunk of stream) {
+  const content = chunk.choices[0]?.delta?.content || '';
+  process.stdout.write(content);
+}
+```
+
+## Error Handling
+
+```python
+from openai import OpenAI, OpenAIError
+
+client = OpenAI(
+    base_url="http://localhost:17765/api/v1",
+    api_key="dummy"
+)
+
+try:
+    response = client.chat.completions.create(
+        model="openai/gpt-4",
+        messages=[
+            {"role": "user", "content": "Hello"}
+        ]
+    )
+    print(response.choices[0].message.content)
+except OpenAIError as e:
+    print(f"Error: {e}")
+```
+
+## Advanced Features
+
+### Context Condensation
+
+When using models with large context windows, AISBF automatically condenses context when approaching limits:
+
+```python
+# Large context will be automatically condensed
+response = client.chat.completions.create(
+    model="gemini/gemini-2.0-flash",
+    messages=[
+        {"role": "user", "content": "Very long prompt..."},
+        # ... many messages
+    ]
+)
+```
+
+### Rate Limiting
+
+AISBF automatically handles rate limits and rotates to available providers:
+
+```python
+# If rate limit is hit, AISBF will automatically use another provider
+for i in range(100):
+    response = client.chat.completions.create(
+        model="coding",  # Rotation with multiple providers
+        messages=[{"role": "user", "content": f"Request {i}"}]
+    )
+```
+
+## Dashboard Access
+
+Access the web dashboard at:
+```
+http://localhost:17765/dashboard
+```
+
+Default credentials:
+- Username: `admin`
+- Password: `admin` (SHA256 hashed in config)
+
+## License
+
+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.

MANIFEST.in

@@ -7,4 +7,7 @@ include aisbf.sh
 include cli.py
 recursive-include config *.json
 recursive-include config *.md
-recursive-include aisbf *.py
+recursive-include aisbf *.py
\ No newline at end of file
+recursive-include templates *.html
+recursive-include templates *.css
+recursive-include templates *.js
\ No newline at end of file

README.md

@@ -109,9 +109,26 @@ See `config/providers.json` and `config/rotations.json` for configuration exampl
 ### General Endpoints
 - `GET /` - Server status and provider list (includes providers, rotations, and autoselect)
 
-### Provider Endpoints
+### 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.)
+
+### 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
 
 ### Rotation Endpoints
 - `GET /api/rotations` - List all available rotation configurations
@@ -131,6 +148,17 @@ See `config/providers.json` and `config/rotations.json` for configuration exampl
   - Supports both streaming and non-streaming responses
 - `GET /api/autoselect/models` - List all models across all autoselect configurations
 
+### 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/rotations` - Edit rotations configuration
+- `GET /dashboard/autoselect` - Edit autoselect configuration
+- `GET /dashboard/settings` - Edit server settings
+- `POST /dashboard/restart` - Restart the server
+
 ## Error Handling
 - Rate limiting for failed requests
 - Automatic retry with provider rotation

aisbf.sh

@@ -41,6 +41,37 @@ fi
 # Create log directory if it doesn't exist
 mkdir -p "$LOG_DIR"
 
+# Function to get port from config file
+get_port() {
+    local CONFIG_FILE="$SHARE_DIR/config/aisbf.json"
+    local DEFAULT_PORT=17765
+    
+    # Check if config file exists
+    if [ ! -f "$CONFIG_FILE" ]; then
+        echo "$DEFAULT_PORT"
+        return
+    fi
+    
+    # Try to read port from config using Python
+    local PORT=$(python3 -c "
+import json
+import sys
+try:
+    with open('$CONFIG_FILE', 'r') as f:
+        config = json.load(f)
+        print(config.get('port', $DEFAULT_PORT))
+except:
+    print($DEFAULT_PORT)
+" 2>/dev/null)
+    
+    # Validate port is a number
+    if [[ "$PORT" =~ ^[0-9]+$ ]]; then
+        echo "$PORT"
+    else
+        echo "$DEFAULT_PORT"
+    fi
+}
+
 # Function to create venv if it doesn't exist
 ensure_venv() {
     if [ ! -d "$VENV_DIR" ]; then
@@ -79,15 +110,20 @@ start_server() {
     # Update venv packages silently
     update_venv
     
+    # Get port from config
+    PORT=$(get_port)
+    
     # Activate the virtual environment
     source $VENV_DIR/bin/activate
     
     # Change to share directory where main.py is located
     cd $SHARE_DIR
     
+    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 17765 2>&1 | while IFS= read -r line; do
+    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"
@@ -115,9 +151,14 @@ start_daemon() {
     # Update venv packages silently
     update_venv
     
+    # Get port from config
+    PORT=$(get_port)
+    
+    echo "Starting AISBF on port $PORT in background..."
+    
     # Start in background with nohup and logging
     # Filter out BrokenPipeError logging errors
-    nohup bash -c "source $VENV_DIR/bin/activate && cd $SHARE_DIR && uvicorn main:app --host 127.0.0.1 --port 17765 2>&1 | grep -v '--- Logging error ---' | grep -v 'BrokenPipeError' | grep -v 'Call stack:' | grep -v 'File .*python' | grep -v 'Message:' | grep -v 'Arguments:'" >> "$LOG_DIR/aisbf_stdout.log" 2>&1 &
+    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 &
     PID=$!
     echo $PID > "$PIDFILE"
     echo "AISBF started in background (PID: $PID)"

aisbf/handlers.py

@@ -648,9 +648,438 @@ class RequestHandler:
             await handler.apply_rate_limit()
 
             models = await handler.get_models()
-            return [model.dict() for model in models]
+            
+            # Enhance model information with context window and capabilities
+            enhanced_models = []
+            for model in models:
+                model_dict = model.dict()
+                model_name = model_dict.get('id', '')
+                
+                # Try to find model config in provider config
+                model_config = None
+                if provider_config.models:
+                    for m in provider_config.models:
+                        if m.name == model_name:
+                            model_config = m
+                            break
+                
+                # Add context window information
+                if model_config and hasattr(model_config, 'context_size'):
+                    model_dict['context_window'] = model_config.context_size
+                elif 'context_window' not in model_dict:
+                    # Try to infer from model name or set a default
+                    model_dict['context_window'] = self._infer_context_window(model_name, provider_config.type)
+                
+                # Add capabilities information
+                if model_config and hasattr(model_config, 'capabilities'):
+                    model_dict['capabilities'] = model_config.capabilities
+                elif 'capabilities' not in model_dict:
+                    # Auto-detect capabilities based on model name and provider type
+                    model_dict['capabilities'] = self._detect_capabilities(model_name, provider_config.type)
+                
+                enhanced_models.append(model_dict)
+            
+            return enhanced_models
         except Exception as e:
             raise HTTPException(status_code=500, detail=str(e))
+    
+    def _infer_context_window(self, model_name: str, provider_type: str) -> int:
+        """Infer context window size from model name or provider type"""
+        model_lower = model_name.lower()
+        
+        # Known model patterns
+        if 'gpt-4' in model_lower:
+            if 'turbo' in model_lower or '1106' in model_lower or '0125' in model_lower:
+                return 128000
+            return 8192
+        elif 'gpt-3.5' in model_lower:
+            if 'turbo' in model_lower and ('1106' in model_lower or '0125' in model_lower):
+                return 16385
+            return 4096
+        elif 'claude-3' in model_lower:
+            return 200000
+        elif 'claude-2' in model_lower:
+            return 100000
+        elif 'gemini' in model_lower:
+            if '1.5' in model_lower:
+                return 2000000 if 'pro' in model_lower else 1000000
+            elif '2.0' in model_lower:
+                return 1000000
+            return 32000
+        elif 'llama' in model_lower:
+            if '3' in model_lower:
+                return 128000
+            return 4096
+        elif 'mistral' in model_lower:
+            if 'large' in model_lower:
+                return 32000
+            return 8192
+        
+        # Default based on provider type
+        if provider_type == 'google':
+            return 32000
+        elif provider_type == 'anthropic':
+            return 100000
+        elif provider_type == 'openai':
+            return 8192
+        
+        # Generic default
+        return 4096
+    
+    def _detect_capabilities(self, model_name: str, provider_type: str) -> List[str]:
+        """Auto-detect model capabilities based on model name and provider type"""
+        model_lower = model_name.lower()
+        capabilities = []
+        
+        # Text-to-text (default for most models)
+        if not any(keyword in model_lower for keyword in ['embedding', 'embed', 'whisper', 'tts', 'dall-e', 'stable-diffusion']):
+            capabilities.append('t2t')
+        
+        # Text-to-image generation
+        if any(keyword in model_lower for keyword in ['dall-e', 'dalle', 'stable-diffusion', 'sd-', 'sdxl', 'midjourney', 'imagen', 'flux']):
+            capabilities.append('t2i')
+        
+        # Image-to-image (editing, style transfer)
+        if any(keyword in model_lower for keyword in ['stable-diffusion', 'sd-', 'sdxl', 'controlnet', 'img2img']):
+            capabilities.append('i2i')
+        
+        # Vision/Image understanding (image-to-text)
+        if any(keyword in model_lower for keyword in ['vision', 'gpt-4-turbo', 'gpt-4o', 'claude-3', 'gemini-1.5', 'gemini-2.0', 'gemini-pro-vision', 'llava', 'blip']):
+            capabilities.append('vision')
+            capabilities.append('i2t')
+        
+        # Audio transcription (audio-to-text)
+        if any(keyword in model_lower for keyword in ['whisper', 'transcribe', 'speech-to-text', 'stt']):
+            capabilities.append('transcription')
+            capabilities.append('a2t')
+        
+        # Text-to-speech
+        if any(keyword in model_lower for keyword in ['tts', 'text-to-speech', 'elevenlabs', 'bark', 'tortoise']):
+            capabilities.append('tts')
+            capabilities.append('t2a')
+        
+        # Text-to-video generation
+        if any(keyword in model_lower for keyword in ['sora', 'runway', 'pika', 'text-to-video', 't2v']):
+            capabilities.append('t2v')
+        
+        # Image-to-video generation
+        if any(keyword in model_lower for keyword in ['runway', 'pika', 'img2video', 'i2v']):
+            capabilities.append('i2v')
+        
+        # Video-to-video (editing)
+        if any(keyword in model_lower for keyword in ['runway', 'video-edit', 'v2v']):
+            capabilities.append('v2v')
+        
+        # Video understanding (video-to-text)
+        if any(keyword in model_lower for keyword in ['video-llama', 'video-chat', 'v2t']):
+            capabilities.append('v2t')
+        
+        # Audio-to-audio (music generation, audio processing)
+        if any(keyword in model_lower for keyword in ['musicgen', 'audiogen', 'riffusion', 'a2a']):
+            capabilities.append('a2a')
+        
+        # Text embeddings
+        if any(keyword in model_lower for keyword in ['embedding', 'embed', 'ada-002', 'bge', 'e5', 'instructor']):
+            capabilities.append('embeddings')
+        
+        # Function calling / tool use
+        if any(keyword in model_lower for keyword in ['gpt-4', 'gpt-3.5-turbo', 'claude-3', 'gemini', 'function', 'tool']):
+            capabilities.append('function_calling')
+        
+        # Code generation
+        if any(keyword in model_lower for keyword in ['codex', 'code-', 'starcoder', 'codellama', 'deepseek-coder', 'phind']):
+            capabilities.append('code_generation')
+            capabilities.append('code_completion')
+        
+        # Translation
+        if any(keyword in model_lower for keyword in ['translate', 'translation', 'm2m', 'nllb']):
+            capabilities.append('translation')
+        
+        # Summarization
+        if any(keyword in model_lower for keyword in ['summarize', 'summary', 'bart', 'pegasus']):
+            capabilities.append('summarization')
+        
+        # Classification
+        if any(keyword in model_lower for keyword in ['classifier', 'classification', 'bert-', 'roberta-']):
+            capabilities.append('classification')
+        
+        # Sentiment analysis
+        if any(keyword in model_lower for keyword in ['sentiment', 'emotion']):
+            capabilities.append('sentiment_analysis')
+        
+        # Named Entity Recognition
+        if any(keyword in model_lower for keyword in ['ner', 'entity', 'spacy']):
+            capabilities.append('ner')
+        
+        # Question answering
+        if any(keyword in model_lower for keyword in ['qa', 'question', 'squad']):
+            capabilities.append('question_answering')
+        
+        # Reasoning (chain-of-thought)
+        if any(keyword in model_lower for keyword in ['reasoning', 'cot', 'o1', 'o3']):
+            capabilities.append('reasoning')
+        
+        # Search / RAG
+        if any(keyword in model_lower for keyword in ['search', 'retrieval', 'rag']):
+            capabilities.append('search')
+        
+        # Content moderation
+        if any(keyword in model_lower for keyword in ['moderation', 'safety', 'content-filter']):
+            capabilities.append('moderation')
+        
+        # Fine-tuning support
+        if any(keyword in model_lower for keyword in ['fine-tune', 'finetune', 'ft-']):
+            capabilities.append('fine_tuning')
+        
+        # Multimodal (multiple input/output types)
+        if any(keyword in model_lower for keyword in ['gpt-4o', 'gemini', 'claude-3', 'multimodal', 'mm-']):
+            capabilities.append('multimodal')
+        
+        # OCR (Optical Character Recognition)
+        if any(keyword in model_lower for keyword in ['ocr', 'tesseract', 'paddleocr', 'easyocr']):
+            capabilities.append('ocr')
+        
+        # Image captioning
+        if any(keyword in model_lower for keyword in ['caption', 'blip', 'git-']):
+            capabilities.append('image_captioning')
+        
+        # Object detection
+        if any(keyword in model_lower for keyword in ['yolo', 'detection', 'rcnn', 'detr']):
+            capabilities.append('object_detection')
+        
+        # Segmentation
+        if any(keyword in model_lower for keyword in ['segment', 'sam', 'mask']):
+            capabilities.append('segmentation')
+        
+        # 3D generation
+        if any(keyword in model_lower for keyword in ['3d', 'nerf', 'gaussian', 'mesh']):
+            capabilities.append('3d_generation')
+        
+        # Animation
+        if any(keyword in model_lower for keyword in ['animate', 'motion', 'pose']):
+            capabilities.append('animation')
+        
+        return capabilities
+    
+    async def handle_audio_transcription(self, request: Request, provider_id: str, form_data) -> Dict:
+        """Handle audio transcription requests"""
+        import logging
+        logger = logging.getLogger(__name__)
+        logger.info(f"=== Audio Transcription Handler START ===")
+        
+        provider_config = self.config.get_provider(provider_id)
+        
+        if provider_config.api_key_required:
+            api_key = request.headers.get('Authorization', '').replace('Bearer ', '')
+            if not api_key:
+                raise HTTPException(status_code=401, detail="API key required")
+        else:
+            api_key = None
+        
+        handler = get_provider_handler(provider_id, api_key)
+        
+        if handler.is_rate_limited():
+            raise HTTPException(status_code=503, detail="Provider temporarily unavailable")
+        
+        try:
+            await handler.apply_rate_limit()
+            result = await handler.handle_audio_transcription(form_data)
+            handler.record_success()
+            return result
+        except Exception as e:
+            handler.record_failure()
+            raise HTTPException(status_code=500, detail=str(e))
+    
+    async def handle_text_to_speech(self, request: Request, provider_id: str, request_data: Dict) -> StreamingResponse:
+        """Handle text-to-speech requests"""
+        import logging
+        logger = logging.getLogger(__name__)
+        logger.info(f"=== Text-to-Speech Handler START ===")
+        
+        provider_config = self.config.get_provider(provider_id)
+        
+        if provider_config.api_key_required:
+            api_key = request_data.get('api_key') or request.headers.get('Authorization', '').replace('Bearer ', '')
+            if not api_key:
+                raise HTTPException(status_code=401, detail="API key required")
+        else:
+            api_key = None
+        
+        handler = get_provider_handler(provider_id, api_key)
+        
+        if handler.is_rate_limited():
+            raise HTTPException(status_code=503, detail="Provider temporarily unavailable")
+        
+        try:
+            await handler.apply_rate_limit()
+            result = await handler.handle_text_to_speech(request_data)
+            handler.record_success()
+            return result
+        except Exception as e:
+            handler.record_failure()
+            raise HTTPException(status_code=500, detail=str(e))
+    
+    async def handle_image_generation(self, request: Request, provider_id: str, request_data: Dict) -> Dict:
+        """Handle image generation requests with URL rewriting"""
+        import logging
+        logger = logging.getLogger(__name__)
+        logger.info(f"=== Image Generation Handler START ===")
+        
+        provider_config = self.config.get_provider(provider_id)
+        
+        if provider_config.api_key_required:
+            api_key = request_data.get('api_key') or request.headers.get('Authorization', '').replace('Bearer ', '')
+            if not api_key:
+                raise HTTPException(status_code=401, detail="API key required")
+        else:
+            api_key = None
+        
+        handler = get_provider_handler(provider_id, api_key)
+        
+        if handler.is_rate_limited():
+            raise HTTPException(status_code=503, detail="Provider temporarily unavailable")
+        
+        try:
+            await handler.apply_rate_limit()
+            result = await handler.handle_image_generation(request_data)
+            
+            # Rewrite URLs in the response to point to our proxy
+            result = self._rewrite_content_urls(result, request)
+            
+            handler.record_success()
+            return result
+        except Exception as e:
+            handler.record_failure()
+            raise HTTPException(status_code=500, detail=str(e))
+    
+    async def handle_embeddings(self, request: Request, provider_id: str, request_data: Dict) -> Dict:
+        """Handle embeddings requests"""
+        import logging
+        logger = logging.getLogger(__name__)
+        logger.info(f"=== Embeddings Handler START ===")
+        
+        provider_config = self.config.get_provider(provider_id)
+        
+        if provider_config.api_key_required:
+            api_key = request_data.get('api_key') or request.headers.get('Authorization', '').replace('Bearer ', '')
+            if not api_key:
+                raise HTTPException(status_code=401, detail="API key required")
+        else:
+            api_key = None
+        
+        handler = get_provider_handler(provider_id, api_key)
+        
+        if handler.is_rate_limited():
+            raise HTTPException(status_code=503, detail="Provider temporarily unavailable")
+        
+        try:
+            await handler.apply_rate_limit()
+            result = await handler.handle_embeddings(request_data)
+            handler.record_success()
+            return result
+        except Exception as e:
+            handler.record_failure()
+            raise HTTPException(status_code=500, detail=str(e))
+    
+    def _rewrite_content_urls(self, response: Dict, request: Request) -> Dict:
+        """Rewrite content URLs to point to our proxy endpoint"""
+        import logging
+        import hashlib
+        import json
+        logger = logging.getLogger(__name__)
+        
+        # Get the base URL from the request
+        scheme = request.url.scheme
+        host = request.headers.get('host', request.url.netloc)
+        base_url = f"{scheme}://{host}"
+        
+        # Store URL mappings in a simple in-memory cache (in production, use Redis or similar)
+        if not hasattr(self, '_url_cache'):
+            self._url_cache = {}
+        
+        def rewrite_url(original_url: str) -> str:
+            """Rewrite a single URL"""
+            # Check if URL is already public and accessible
+            if self._is_public_url(original_url):
+                logger.info(f"URL is public, passing through: {original_url}")
+                return original_url
+            
+            # Generate a unique ID for this URL
+            url_hash = hashlib.md5(original_url.encode()).hexdigest()[:16]
+            
+            # Store the mapping
+            self._url_cache[url_hash] = original_url
+            
+            # Return the proxy URL
+            proxy_url = f"{base_url}/api/proxy/{url_hash}"
+            logger.info(f"Rewrote URL: {original_url} -> {proxy_url}")
+            return proxy_url
+        
+        # Recursively rewrite URLs in the response
+        def rewrite_recursive(obj):
+            if isinstance(obj, dict):
+                for key, value in obj.items():
+                    if key in ['url', 'image_url', 'audio_url', 'video_url'] and isinstance(value, str):
+                        obj[key] = rewrite_url(value)
+                    else:
+                        obj[key] = rewrite_recursive(value)
+            elif isinstance(obj, list):
+                return [rewrite_recursive(item) for item in obj]
+            return obj
+        
+        return rewrite_recursive(response)
+    
+    def _is_public_url(self, url: str) -> bool:
+        """Check if a URL is publicly accessible (doesn't need proxying)"""
+        # URLs from major CDNs and public services don't need proxying
+        public_domains = [
+            'cloudflare.com',
+            'amazonaws.com',
+            'googleusercontent.com',
+            'azure.com',
+            'cdn.',
+            'storage.googleapis.com'
+        ]
+        
+        return any(domain in url.lower() for domain in public_domains)
+    
+    async def handle_content_proxy(self, content_id: str) -> StreamingResponse:
+        """Proxy content from the original URL"""
+        import logging
+        import httpx
+        logger = logging.getLogger(__name__)
+        
+        # Get the original URL from cache
+        if not hasattr(self, '_url_cache'):
+            self._url_cache = {}
+        
+        original_url = self._url_cache.get(content_id)
+        if not original_url:
+            raise HTTPException(status_code=404, detail="Content not found")
+        
+        logger.info(f"Proxying content: {content_id} -> {original_url}")
+        
+        try:
+            # Fetch the content from the original URL
+            async with httpx.AsyncClient() as client:
+                response = await client.get(original_url, follow_redirects=True)
+                response.raise_for_status()
+                
+                # Determine content type
+                content_type = response.headers.get('content-type', 'application/octet-stream')
+                
+                # Return the content as a streaming response
+                return StreamingResponse(
+                    iter([response.content]),
+                    media_type=content_type,
+                    headers={
+                        'Content-Disposition': response.headers.get('content-disposition', ''),
+                        'Cache-Control': 'public, max-age=3600'
+                    }
+                )
+        except Exception as e:
+            logger.error(f"Error proxying content: {str(e)}")
+            raise HTTPException(status_code=500, detail=f"Error fetching content: {str(e)}")
 
 class RotationHandler:
     def __init__(self):
@@ -1940,16 +2369,127 @@ class RotationHandler:
 
         all_models = []
         for provider in rotation_config.providers:
+            provider_id = provider['provider_id']
+            provider_config = self.config.get_provider(provider_id)
+            
             for model in provider['models']:
-                all_models.append({
-                    "id": f"{provider['provider_id']}/{model['name']}",
-                    "name": model['name'],
-                    "provider_id": provider['provider_id'],
+                model_name = model['name']
+                model_dict = {
+                    "id": f"{provider_id}/{model_name}",
+                    "name": model_name,
+                    "provider_id": provider_id,
                     "weight": model['weight'],
                     "rate_limit": model.get('rate_limit')
-                })
+                }
+                
+                # Add context window information
+                if model.get('context_size'):
+                    model_dict['context_window'] = model['context_size']
+                elif provider_config:
+                    # Try to find in provider config
+                    for pm in provider_config.models or []:
+                        if pm.name == model_name and hasattr(pm, 'context_size'):
+                            model_dict['context_window'] = pm.context_size
+                            break
+                    if 'context_window' not in model_dict:
+                        model_dict['context_window'] = self._infer_context_window(model_name, provider_config.type)
+                
+                # Add capabilities information
+                if model.get('capabilities'):
+                    model_dict['capabilities'] = model['capabilities']
+                elif provider_config:
+                    # Try to find in provider config
+                    for pm in provider_config.models or []:
+                        if pm.name == model_name and hasattr(pm, 'capabilities'):
+                            model_dict['capabilities'] = pm.capabilities
+                            break
+                    if 'capabilities' not in model_dict:
+                        model_dict['capabilities'] = self._detect_capabilities(model_name, provider_config.type)
+                
+                all_models.append(model_dict)
 
         return all_models
+    
+    def _infer_context_window(self, model_name: str, provider_type: str) -> int:
+        """Infer context window size from model name or provider type"""
+        model_lower = model_name.lower()
+        
+        # Known model patterns
+        if 'gpt-4' in model_lower:
+            if 'turbo' in model_lower or '1106' in model_lower or '0125' in model_lower:
+                return 128000
+            return 8192
+        elif 'gpt-3.5' in model_lower:
+            if 'turbo' in model_lower and ('1106' in model_lower or '0125' in model_lower):
+                return 16385
+            return 4096
+        elif 'claude-3' in model_lower:
+            return 200000
+        elif 'claude-2' in model_lower:
+            return 100000
+        elif 'gemini' in model_lower:
+            if '1.5' in model_lower:
+                return 2000000 if 'pro' in model_lower else 1000000
+            elif '2.0' in model_lower:
+                return 1000000
+            return 32000
+        elif 'llama' in model_lower:
+            if '3' in model_lower:
+                return 128000
+            return 4096
+        elif 'mistral' in model_lower:
+            if 'large' in model_lower:
+                return 32000
+            return 8192
+        
+        # Default based on provider type
+        if provider_type == 'google':
+            return 32000
+        elif provider_type == 'anthropic':
+            return 100000
+        elif provider_type == 'openai':
+            return 8192
+        
+        # Generic default
+        return 4096
+    
+    def _detect_capabilities(self, model_name: str, provider_type: str) -> List[str]:
+        """Auto-detect model capabilities based on model name and provider type"""
+        model_lower = model_name.lower()
+        capabilities = []
+        
+        # Text-to-text is the default capability for all models
+        capabilities.append('t2t')
+        
+        # Image generation models
+        if any(keyword in model_lower for keyword in ['dall-e', 'dalle', 'stable-diffusion', 'sd-', 'midjourney', 'imagen']):
+            capabilities.append('t2i')
+        
+        # Vision models (can process images)
+        if any(keyword in model_lower for keyword in ['vision', 'gpt-4-turbo', 'gpt-4o', 'claude-3', 'gemini-1.5', 'gemini-2.0']):
+            capabilities.append('vision')
+        
+        # Audio transcription models
+        if any(keyword in model_lower for keyword in ['whisper', 'transcribe']):
+            capabilities.append('transcription')
+        
+        # Text-to-speech models
+        if any(keyword in model_lower for keyword in ['tts', 'text-to-speech', 'elevenlabs']):
+            capabilities.append('tts')
+        
+        # Video generation models
+        if any(keyword in model_lower for keyword in ['sora', 'runway', 'pika', 'video']):
+            capabilities.append('i2v')
+        
+        # Embedding models
+        if any(keyword in model_lower for keyword in ['embedding', 'embed', 'ada-002']):
+            capabilities.append('embeddings')
+        
+        # Function calling / tool use
+        if any(keyword in model_lower for keyword in ['gpt-4', 'gpt-3.5-turbo', 'claude-3', 'gemini']):
+            capabilities.append('function_calling')
+        
+        return capabilities
 
 class AutoselectHandler:
     def __init__(self):

aisbf/providers.py

@@ -49,6 +49,180 @@ class BaseProviderHandler:
         self.model_last_request_time = {}  # {model_name: timestamp}
         # Token usage tracking for rate limits
         self.token_usage = {}  # {model_name: {"TPM": [], "TPH": [], "TPD": []}}
+    
+    def parse_429_response(self, response_data: Union[Dict, str], headers: Dict = None) -> Optional[int]:
+        """
+        Parse 429 rate limit response to extract wait time in seconds.
+        
+        Checks multiple sources:
+        1. Retry-After header (seconds or HTTP date)
+        2. X-RateLimit-Reset header (Unix timestamp)
+        3. Response body fields (retry_after, reset_time, etc.)
+        
+        Returns:
+            Wait time in seconds, or None if cannot be determined
+        """
+        import logging
+        import re
+        from email.utils import parsedate_to_datetime
+        from datetime import datetime, timezone
+        
+        logger = logging.getLogger(__name__)
+        logger.info("=== PARSING 429 RATE LIMIT RESPONSE ===")
+        
+        wait_seconds = None
+        
+        # Check Retry-After header
+        if headers:
+            retry_after = headers.get('Retry-After') or headers.get('retry-after')
+            if retry_after:
+                logger.info(f"Found Retry-After header: {retry_after}")
+                try:
+                    # Try parsing as integer (seconds)
+                    wait_seconds = int(retry_after)
+                    logger.info(f"Parsed Retry-After as seconds: {wait_seconds}")
+                except ValueError:
+                    # Try parsing as HTTP date
+                    try:
+                        retry_date = parsedate_to_datetime(retry_after)
+                        now = datetime.now(timezone.utc)
+                        wait_seconds = int((retry_date - now).total_seconds())
+                        logger.info(f"Parsed Retry-After as date, wait seconds: {wait_seconds}")
+                    except Exception as e:
+                        logger.warning(f"Failed to parse Retry-After header: {e}")
+            
+            # Check X-RateLimit-Reset header (Unix timestamp)
+            if not wait_seconds:
+                reset_time = headers.get('X-RateLimit-Reset') or headers.get('x-ratelimit-reset')
+                if reset_time:
+                    logger.info(f"Found X-RateLimit-Reset header: {reset_time}")
+                    try:
+                        reset_timestamp = int(reset_time)
+                        now_timestamp = int(time.time())
+                        wait_seconds = reset_timestamp - now_timestamp
+                        logger.info(f"Calculated wait from reset timestamp: {wait_seconds} seconds")
+                    except Exception as e:
+                        logger.warning(f"Failed to parse X-RateLimit-Reset header: {e}")
+        
+        # Check response body
+        if not wait_seconds and isinstance(response_data, dict):
+            logger.info(f"Checking response body for rate limit info: {response_data}")
+            
+            # Common field names for retry/reset time
+            retry_fields = [
+                'retry_after', 'retryAfter', 'retry_after_seconds',
+                'wait_seconds', 'waitSeconds', 'retry_in'
+            ]
+            reset_fields = [
+                'reset_time', 'resetTime', 'reset_at', 'resetAt',
+                'reset_timestamp', 'resetTimestamp'
+            ]
+            
+            # Check retry fields (direct seconds)
+            for field in retry_fields:
+                if field in response_data:
+                    try:
+                        wait_seconds = int(response_data[field])
+                        logger.info(f"Found {field} in response body: {wait_seconds} seconds")
+                        break
+                    except (ValueError, TypeError) as e:
+                        logger.warning(f"Failed to parse {field}: {e}")
+            
+            # Check reset fields (timestamp)
+            if not wait_seconds:
+                for field in reset_fields:
+                    if field in response_data:
+                        try:
+                            reset_timestamp = int(response_data[field])
+                            now_timestamp = int(time.time())
+                            wait_seconds = reset_timestamp - now_timestamp
+                            logger.info(f"Found {field} in response body, calculated wait: {wait_seconds} seconds")
+                            break
+                        except (ValueError, TypeError) as e:
+                            logger.warning(f"Failed to parse {field}: {e}")
+            
+            # Check for error message with time information
+            if not wait_seconds:
+                error_msg = response_data.get('error', {})
+                if isinstance(error_msg, dict):
+                    message = error_msg.get('message', '')
+                elif isinstance(error_msg, str):
+                    message = error_msg
+                else:
+                    message = response_data.get('message', '')
+                
+                if message:
+                    logger.info(f"Checking error message for time info: {message}")
+                    # Look for patterns like "try again in X seconds/minutes/hours"
+                    patterns = [
+                        r'try again in (\d+)\s*(second|minute|hour|day)s?',
+                        r'retry after (\d+)\s*(second|minute|hour|day)s?',
+                        r'wait (\d+)\s*(second|minute|hour|day)s?',
+                        r'available in (\d+)\s*(second|minute|hour|day)s?',
+                    ]
+                    
+                    for pattern in patterns:
+                        match = re.search(pattern, message, re.IGNORECASE)
+                        if match:
+                            value = int(match.group(1))
+                            unit = match.group(2).lower()
+                            
+                            # Convert to seconds
+                            multipliers = {
+                                'second': 1,
+                                'minute': 60,
+                                'hour': 3600,
+                                'day': 86400
+                            }
+                            wait_seconds = value * multipliers.get(unit, 1)
+                            logger.info(f"Extracted wait time from message: {value} {unit}(s) = {wait_seconds} seconds")
+                            break
+        
+        # Ensure wait_seconds is positive and reasonable
+        if wait_seconds:
+            if wait_seconds < 0:
+                logger.warning(f"Calculated negative wait time: {wait_seconds}, setting to 60 seconds")
+                wait_seconds = 60
+            elif wait_seconds > 86400:  # More than 1 day
+                logger.warning(f"Calculated very long wait time: {wait_seconds}, capping at 1 day")
+                wait_seconds = 86400
+            
+            logger.info(f"Final parsed wait time: {wait_seconds} seconds")
+        else:
+            logger.warning("Could not determine wait time from 429 response, using default 60 seconds")
+            wait_seconds = 60
+        
+        logger.info("=== END PARSING 429 RATE LIMIT RESPONSE ===")
+        return wait_seconds
+    
+    def handle_429_error(self, response_data: Union[Dict, str] = None, headers: Dict = None):
+        """
+        Handle 429 rate limit error by parsing the response and disabling provider
+        for the appropriate duration.
+        
+        Args:
+            response_data: Response body (dict or string)
+            headers: Response headers
+        """
+        import logging
+        logger = logging.getLogger(__name__)
+        
+        logger.error("=== 429 RATE LIMIT ERROR DETECTED ===")
+        logger.error(f"Provider: {self.provider_id}")
+        
+        # Parse the response to get wait time
+        wait_seconds = self.parse_429_response(response_data, headers)
+        
+        # Disable provider for the calculated duration
+        self.error_tracking['disabled_until'] = time.time() + wait_seconds
+        
+        logger.error(f"!!! PROVIDER DISABLED DUE TO RATE LIMIT !!!")
+        logger.error(f"Provider: {self.provider_id}")
+        logger.error(f"Reason: 429 Too Many Requests")
+        logger.error(f"Disabled for: {wait_seconds} seconds ({wait_seconds / 60:.1f} minutes)")
+        logger.error(f"Disabled until: {self.error_tracking['disabled_until']}")
+        logger.error(f"Provider will be automatically re-enabled after cooldown")
+        logger.error("=== END 429 RATE LIMIT ERROR ===")
 
     def is_rate_limited(self) -> bool:
         if self.error_tracking['disabled_until'] and self.error_tracking['disabled_until'] > time.time():
@@ -1329,6 +1503,19 @@ class KiroProviderHandler(BaseProviderHandler):
                 headers=headers
             )
             
+            # Check for 429 rate limit error before raising
+            if response.status_code == 429:
+                try:
+                    response_data = response.json()
+                except Exception:
+                    response_data = response.text
+                
+                # Handle 429 error with intelligent parsing
+                self.handle_429_error(response_data, dict(response.headers))
+                
+                # Re-raise the error after handling
+                response.raise_for_status()
+            
             response.raise_for_status()
             response_data = response.json()
             
@@ -1534,6 +1721,20 @@ class OllamaProviderHandler(BaseProviderHandler):
             logger.info(f"Response content type: {response.headers.get('content-type')}")
             logger.info(f"Response content length: {len(response.content)} bytes")
             logger.info(f"Raw response content (first 500 chars): {response.text[:500]}")
+            
+            # Check for 429 rate limit error before raising
+            if response.status_code == 429:
+                try:
+                    response_data = response.json()
+                except Exception:
+                    response_data = response.text
+                
+                # Handle 429 error with intelligent parsing
+                self.handle_429_error(response_data, dict(response.headers))
+                
+                # Re-raise the error after handling
+                response.raise_for_status()
+            
             response.raise_for_status()
             
             # Ollama may return multiple JSON objects, parse them all

config/aisbf.json

@@ -16,7 +16,7 @@
   "dashboard": {
     "enabled": true,
     "username": "admin",
-    "password": "admin"
+    "password": "8c6976e5b5410415bde908bd4dee15dfb167a9c873fc4bb8a81f6f2ab448a918"
   },
   "internal_model": {
     "model_id": "huihui-ai/Qwen2.5-0.5B-Instruct-abliterated-v3"

config/providers.json

@@ -24,7 +24,8 @@
           "rate_limit_TPD": 1000000,
           "context_size": 1000000,
           "condense_context": 80,
-          "condense_method": ["hierarchical", "semantic"]
+          "condense_method": ["hierarchical", "semantic"],
+          "capabilities": ["t2t", "vision", "function_calling"]
         },
         {
           "name": "gemini-1.5-pro",
@@ -35,7 +36,8 @@
           "rate_limit_TPD": 1000000,
           "context_size": 2000000,
           "condense_context": 85,
-          "condense_method": "conversational"
+          "condense_method": "conversational",
+          "capabilities": ["t2t", "vision", "function_calling"]
         }
       ]
     },

main.py

@@ -38,6 +38,7 @@ import sys
 import os
 import argparse
 import secrets
+import hashlib
 from logging.handlers import RotatingFileHandler
 from datetime import datetime, timedelta
 from collections import defaultdict
@@ -47,6 +48,9 @@ import json
 # Global variable to store custom config directory
 _custom_config_dir = None
 
+# Global variable to store original command line arguments for restart
+_original_argv = None
+
 def set_config_dir(config_dir: str):
     """Set custom config directory before importing config"""
     global _custom_config_dir
@@ -449,7 +453,14 @@ async def dashboard_login_page(request: Request):
 async def dashboard_login(request: Request, username: str = Form(...), password: str = Form(...)):
     """Handle dashboard login"""
     dashboard_config = server_config.get('dashboard_config', {})
-    if username == dashboard_config.get('username', 'admin') and password == dashboard_config.get('password', 'admin'):
+    stored_username = dashboard_config.get('username', 'admin')
+    stored_password_hash = dashboard_config.get('password', '8c6976e5b5410415bde908bd4dee15dfb167a9c873fc4bb8a81f6f2ab448a918')
+    
+    # Hash the submitted password
+    password_hash = hashlib.sha256(password.encode()).hexdigest()
+    
+    # Compare username and hashed 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)
@@ -735,8 +746,9 @@ async def dashboard_settings_save(
     aisbf_config['auth']['enabled'] = auth_enabled
     aisbf_config['auth']['tokens'] = [t.strip() for t in auth_tokens.split('\n') if t.strip()]
     aisbf_config['dashboard']['username'] = dashboard_username
-    if dashboard_password:  # Only update if provided
-        aisbf_config['dashboard']['password'] = dashboard_password
+    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
     
     # Save config
@@ -752,6 +764,47 @@ async def dashboard_settings_save(
         "success": "Settings saved successfully! Restart server for changes to take effect."
     })
 
+@app.post("/dashboard/restart")
+async def dashboard_restart(request: Request):
+    """Restart the server"""
+    auth_check = require_dashboard_auth(request)
+    if auth_check:
+        return auth_check
+    
+    import os
+    import signal
+    
+    logger.info("Server restart requested from dashboard")
+    
+    # Schedule restart after response is sent
+    def restart_server():
+        import time
+        time.sleep(1)  # Give time for response to be sent
+        logger.info("Restarting server...")
+        os.execv(sys.executable, [sys.executable] + _original_argv)
+    
+    import threading
+    threading.Thread(target=restart_server, daemon=True).start()
+    
+    return JSONResponse({"message": "Server is restarting..."})
+
+def parse_provider_from_model(model: str) -> tuple[str, str]:
+    """
+    Parse provider and model from model field.
+    
+    Supports formats:
+    - "provider/model" -> ("provider", "model")
+    - "provider/namespace/model" -> ("provider", "namespace/model")
+    - "model" -> (None, "model")
+    
+    Returns:
+        tuple: (provider_id, actual_model_name)
+    """
+    if '/' in model:
+        parts = model.split('/', 1)
+        return parts[0], parts[1]
+    return None, model
+
 @app.get("/")
 async def root():
     return {
@@ -761,6 +814,173 @@ async def root():
         "autoselect": list(config.autoselect.keys())
     }
 
+# Standard OpenAI-compatible v1 endpoints
+@app.post("/api/v1/chat/completions")
+async def v1_chat_completions(request: Request, body: ChatCompletionRequest):
+    """Standard OpenAI-compatible chat completions endpoint"""
+    logger.info(f"=== V1 CHAT COMPLETION REQUEST ===")
+    logger.info(f"Model: {body.model}")
+    
+    # Parse provider from model field
+    provider_id, actual_model = parse_provider_from_model(body.model)
+    
+    if not provider_id:
+        raise HTTPException(
+            status_code=400,
+            detail="Model must be in format 'provider/model' (e.g., 'openai/gpt-4')"
+        )
+    
+    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:
+        if body.stream:
+            return await autoselect_handler.handle_autoselect_streaming_request(provider_id, body_dict)
+        else:
+            return await autoselect_handler.handle_autoselect_request(provider_id, 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)
+    
+    # Check if it's a provider
+    if provider_id not in config.providers:
+        raise HTTPException(
+            status_code=400,
+            detail=f"Provider '{provider_id}' not found. Available: {list(config.providers.keys())}"
+        )
+    
+    # Handle as direct provider request
+    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 ===")
+    
+    all_models = []
+    
+    # Add provider models
+    for provider_id in config.providers.keys():
+        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)
+        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():
+        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)
+        except Exception as e:
+            logger.warning(f"Error listing models for rotation {rotation_id}: {e}")
+    
+    # Add autoselect models
+    for autoselect_id in config.autoselect.keys():
+        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)
+        except Exception as e:
+            logger.warning(f"Error listing models for autoselect {autoselect_id}: {e}")
+    
+    return {"object": "list", "data": all_models}
+
+@app.post("/api/v1/audio/transcriptions")
+async def v1_audio_transcriptions(request: Request):
+    """Standard audio transcription endpoint"""
+    logger.info("=== V1 AUDIO TRANSCRIPTION REQUEST ===")
+    
+    form = await request.form()
+    model = form.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' (e.g., 'openai/whisper-1')"
+        )
+    
+    # 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
+    
+    return await request_handler.handle_audio_transcription(request, provider_id, updated_form)
+
+@app.post("/api/v1/audio/speech")
+async def v1_audio_speech(request: Request, body: dict):
+    """Standard text-to-speech endpoint"""
+    logger.info("=== V1 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' (e.g., 'openai/tts-1')"
+        )
+    
+    body['model'] = actual_model
+    return await request_handler.handle_text_to_speech(request, provider_id, body)
+
+@app.post("/api/v1/images/generations")
+async def v1_image_generations(request: Request, body: dict):
+    """Standard image generation endpoint"""
+    logger.info("=== V1 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' (e.g., 'openai/dall-e-3')"
+        )
+    
+    body['model'] = actual_model
+    return await request_handler.handle_image_generation(request, provider_id, body)
+
+@app.post("/api/v1/embeddings")
+async def v1_embeddings(request: Request, body: dict):
+    """Standard embeddings endpoint"""
+    logger.info("=== V1 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' (e.g., 'openai/text-embedding-ada-002')"
+        )
+    
+    body['model'] = actual_model
+    return await request_handler.handle_embeddings(request, provider_id, body)
+
 @app.get("/api/rotations")
 async def list_rotations():
     """List all available rotations"""
@@ -1034,6 +1254,78 @@ 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}")
+    
+    # Get form data (audio file upload)
+    form = await request.form()
+    
+    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))
+
+@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}")
+    
+    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))
+
+# 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}")
+    
+    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))
+
+# 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}")
+    
+    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))
+
+# Content proxy endpoint
+@app.get("/api/proxy/{content_id}")
+async def proxy_content(content_id: str):
+    """Proxy generated content (images, audio, etc.)"""
+    logger.info(f"=== PROXY CONTENT REQUEST ===")
+    logger.info(f"Content ID: {content_id}")
+    
+    try:
+        result = await request_handler.handle_content_proxy(content_id)
+        return result
+    except Exception as e:
+        logger.error(f"Error proxying content: {str(e)}", exc_info=True)
+        raise HTTPException(status_code=500, detail=str(e))
+
 @app.post("/api/{provider_id}")
 async def catch_all_post(provider_id: str, request: Request):
     """Catch-all for POST requests to help debug routing issues"""
@@ -1086,6 +1378,10 @@ Examples:
     
     args = parser.parse_args()
     
+    # Store original command line arguments for restart functionality
+    global _original_argv
+    _original_argv = sys.argv.copy()
+    
     # Set custom config directory if provided
     if args.config:
         set_config_dir(args.config)

setup.py

@@ -86,6 +86,7 @@ setup(
             'config/rotations.json',
             'config/autoselect.json',
             'config/autoselect.md',
+            'config/aisbf.json',
         ]),
         # Install aisbf package to share directory for venv installation
         ('share/aisbf/aisbf', [
@@ -98,6 +99,16 @@ setup(
             'aisbf/utils.py',
             'aisbf/database.py',
         ]),
+        # Install dashboard templates
+        ('share/aisbf/templates', [
+            'templates/base.html',
+        ]),
+        ('share/aisbf/templates/dashboard', [
+            'templates/dashboard/login.html',
+            'templates/dashboard/index.html',
+            'templates/dashboard/edit_config.html',
+            'templates/dashboard/settings.html',
+        ]),
     ],
     entry_points={
         "console_scripts": [

templates/base.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/>.
+-->
 <!DOCTYPE html>
 <html lang="en">
 <head>
@@ -9,7 +25,8 @@
         body { font-family: -apple-system, BlinkMacSystemFont, 'Segoe UI', Roboto, Oxygen, Ubuntu, Cantarell, sans-serif; background: #f5f5f5; }
         .container { max-width: 1200px; margin: 0 auto; padding: 20px; }
         .header { background: #2c3e50; color: white; padding: 20px 0; margin-bottom: 30px; }
-        .header h1 { font-size: 24px; font-weight: 600; }
+        .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; }
@@ -19,28 +36,67 @@
         .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 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; }
+        .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-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; }
-        .logout { float: right; }
+        .alert-info { background: #d1ecf1; color: #0c5460; border: 1px solid #bee5eb; }
         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; }
     </style>
+    <script>
+        /*
+         * 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/>.
+         */
+        function restartServer() {
+            if (confirm('Are you sure you want to restart the server? This will disconnect all active connections.')) {
+                fetch('/dashboard/restart', {
+                    method: 'POST',
+                    headers: {'Content-Type': 'application/json'}
+                })
+                .then(response => response.json())
+                .then(data => {
+                    alert(data.message + ' The page will reload in 5 seconds.');
+                    setTimeout(() => window.location.reload(), 5000);
+                })
+                .catch(error => {
+                    alert('Error restarting server: ' + error);
+                });
+            }
+        }
+    </script>
 </head>
 <body>
     <div class="header">
         <div class="container">
             <h1>AISBF Dashboard</h1>
             {% if session.logged_in %}
-            <a href="/dashboard/logout" class="btn btn-secondary logout">Logout</a>
+            <div class="header-actions">
+                <button onclick="restartServer()" class="btn btn-warning">Restart Server</button>
+                <a href="/dashboard/logout" class="btn btn-secondary">Logout</a>
+            </div>
             {% endif %}
         </div>
     </div>

templates/dashboard/edit_config.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 %}

templates/dashboard/index.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 %}Overview - AISBF Dashboard{% endblock %}

templates/dashboard/login.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 %}Login - AISBF Dashboard{% endblock %}

templates/dashboard/settings.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 %}Settings - AISBF Dashboard{% endblock %}