YT_Skip is an AI-powered tool that automatically identifies and skips unnecessary parts of YouTube videos using cutting-edge language models. Powered by Groq's ultra-fast inference and Meta's Llama 4 Scout model, it delivers lightning-speed analysis to enhance your viewing experience by removing filler content, advertisements, and redundant explanations.
- ⚡ Ultra-Fast Processing: Powered by Groq's high-speed inference infrastructure
- 🧠 Advanced AI: Uses Meta's Llama 4 Scout 17B model with 128K context window
- 🎯 Precision Skipping: Enhanced confidence scoring for more accurate skip detection
- 📊 Smart Analysis: Adaptive prompting based on video characteristics
- 🔍 Detailed Categorization: Skip segments are classified by reason (ads, filler, etc.)
- 💾 Intelligent Caching: Reduces processing time for repeat requests
- 🎨 User Preferences: Customizable skip categories and sensitivity levels
- 🛡️ Robust Error Handling: Comprehensive error responses and edge case management
- Automatic Detection: AI-powered identification of non-essential video segments
- Smart Categorization: Classifies skips as advertisements, filler speech, repetitive content, etc.
- User Preferences: 10+ predefined categories plus custom keywords/phrases
- Confidence Scoring: Each skip segment includes a confidence score
- Real-time Processing: Ultra-fast response times with Groq infrastructure
- Caching System: Efficient caching to minimize API calls
- RESTful API: Easy integration with Chrome extensions and other tools
- Health Monitoring: Built-in health checks and usage statistics
- Backend: FastAPI with Python 3.10
- AI Provider: Groq Cloud (ultra-fast inference)
- Model: Meta Llama 4 Scout 17B 16E Instruct
- Transcript Source: YouTube Transcript API
- Containerization: Docker with Docker Compose
- Docker and Docker Compose
- Groq API Key (Get one free here)
-
Clone the repository:
git clone <repository-url> cd YouTube-Transcript-Summarizer
-
Set your Groq API key in
docker-compose.yml:environment: - GROQ_API_KEY=your_groq_api_key
-
Run with Docker Compose:
docker-compose up --build
-
Test the integration:
export GROQ_API_KEY=your_api_key python test_groq.py
pip install -r requirements.txt
export GROQ_API_KEY=your_groq_api_key
uvicorn backend.app:app --reload --host 0.0.0.0 --port 8000http://localhost:8000
All endpoints require the GROQ_API_KEY to be set in the environment. No additional authentication is needed for individual requests.
Endpoint: GET /process_video
Description: Process a YouTube video to identify skip segments with optional user preferences.
Parameters:
video_id(string, required): YouTube video ID (11 characters)user_preferences(UserPreferences, optional): Skip preferences object
Example Request:
curl -X GET "http://localhost:8000/process_video?video_id=dQw4w9WgXcQ"Example Response:
{
"transcription": [
{
"text": "Welcome back to my channel",
"start": 0.0,
"duration": 2.5
}
],
"remove": [
{
"start": 15.2,
"end": 18.7,
"confidence": 0.85,
"reason": "Call to Action"
}
],
"processing_time": 1.23,
"total_duration": 180.5,
"skip_percentage": 12.4
}Endpoint: POST /process_video
Description: Process a YouTube video with detailed user preferences via POST body.
Request Body:
{
"video_id": "dQw4w9WgXcQ",
"user_preferences": {
"default_categories": ["advertisements", "calls_to_action"],
"custom_keywords": ["crypto", "NFT"],
"custom_phrases": ["my personal opinion"],
"sensitivity": "medium",
"enabled": true
}
}Example Request:
curl -X POST "http://localhost:8000/process_video" \
-H "Content-Type: application/json" \
-d '{
"video_id": "dQw4w9WgXcQ",
"user_preferences": {
"default_categories": ["advertisements"],
"custom_keywords": ["sponsor"],
"sensitivity": "high",
"enabled": true
}
}'Response: Same as GET endpoint
Endpoint: GET /health
Description: Check service health and cache status.
Example Request:
curl -X GET "http://localhost:8000/health"Example Response:
{
"status": "healthy",
"cache_size": 42,
"model": "meta-llama/llama-4-scout-17b-16e-instruct",
"provider": "Groq"
}Endpoint: GET /api/stats
Description: Get detailed API usage statistics and model information.
Example Request:
curl -X GET "http://localhost:8000/api/stats"Example Response:
{
"total_cached_videos": 42,
"model_info": {
"name": "meta-llama/llama-4-scout-17b-16e-instruct",
"provider": "Groq",
"context_window": "128K tokens",
"features": ["ultra-fast inference", "multimodal", "JSON mode"]
}
}Endpoint: DELETE /cache/{video_id}
Description: Clear cache entries for a specific video.
Parameters:
video_id(string, required): YouTube video ID
Example Request:
curl -X DELETE "http://localhost:8000/cache/dQw4w9WgXcQ"Example Response:
{
"message": "Cleared 3 cache entries for video dQw4w9WgXcQ"
}Endpoint: GET /preferences/categories
Description: Get all available default skip categories with descriptions.
Example Request:
curl -X GET "http://localhost:8000/preferences/categories"Example Response:
{
"categories": {
"advertisements": {
"name": "Advertisements",
"description": "Keywords: sponsor, sponsored, ad...",
"keyword_count": 9,
"phrase_count": 5
}
},
"total_categories": 10
}Endpoint: GET /preferences/category/{category_name}
Description: Get detailed information about a specific skip category.
Parameters:
category_name(string, required): Category name (e.g., "advertisements")
Example Request:
curl -X GET "http://localhost:8000/preferences/category/advertisements"Example Response:
{
"name": "Advertisements",
"keywords": ["sponsor", "sponsored", "ad", "advertisement"],
"phrases": ["this video is sponsored by", "today's sponsor"],
"total_terms": 14
}{
"default_categories": ["string"], // Available: advertisements, calls_to_action, etc.
"custom_keywords": ["string"], // Custom words to skip
"custom_phrases": ["string"], // Custom phrases to skip
"sensitivity": "string", // "low" | "medium" | "high"
"enabled": true // Enable/disable preferences
}{
"start": 15.2, // Start time in seconds
"end": 18.7, // End time in seconds
"confidence": 0.85, // Confidence score (0.0-1.0)
"reason": "Advertisement" // Classification reason
}{
"text": "Hello everyone", // Transcript text
"start": 0.0, // Start time in seconds
"duration": 2.5 // Duration in seconds
}{
"detail": "Transcripts are disabled for this video."
}{
"detail": "No transcript found for this video."
}{
"detail": "Category not found"
}{
"detail": "Error processing with Groq: API rate limit exceeded"
}- Issue: Transcript exceeds token limits
- Handling: Smart chunking preserves beginning, middle samples, and end
- Result: Maintains quality while fitting context window
- Issue: Auto-generated transcripts disabled or unavailable
- Handling: Returns 400 error with clear message
- Solution: Check video has captions before processing
- Issue: Minimal content to analyze
- Handling: Reduced sensitivity, preserves most content
- Result: Conservative skipping to maintain value
- Issue: AI model optimized for English
- Handling: Attempts processing but may have reduced accuracy
- Solution: Works best with English transcripts
- Issue: Groq API rate limits exceeded
- Handling: Returns 500 error with specific message
- Solution: Implement backoff strategy or upgrade plan
- Issue: Stale cache entries
- Handling: 24-hour automatic expiry
- Manual: Use DELETE /cache/{video_id} endpoint
- Issue: Malformed or non-existent video IDs
- Handling: YouTube API validation before processing
- Result: Clear error message for invalid IDs
- Issue: Large video processing
- Handling: Streaming processing and garbage collection
- Result: Efficient memory usage
// Example error handling in client code
async function processVideo(videoId, preferences) {
try {
const response = await fetch('/process_video', {
method: 'POST',
headers: { 'Content-Type': 'application/json' },
body: JSON.stringify({ video_id: videoId, user_preferences: preferences })
});
if (!response.ok) {
const error = await response.json();
// Handle specific error cases
switch (response.status) {
case 400:
if (error.detail.includes('transcript')) {
showMessage('This video has no available transcript');
return null;
}
break;
case 500:
if (error.detail.includes('rate limit')) {
showMessage('Service temporarily busy, please try again');
return null;
}
break;
}
throw new Error(error.detail);
}
return await response.json();
} catch (error) {
console.error('Processing failed:', error);
throw error;
}
}GROQ_API_KEY=your_groq_api_key # Required: Groq API key
DEV_MODE=true # Optional: Enable development mode- Model:
meta-llama/llama-4-scout-17b-16e-instruct - Temperature: 0.1 (for consistency)
- Max Tokens: 2048
- Context Window: 128K tokens
CACHE_EXPIRY_HOURS = 24 # Cache entries expire after 24 hoursadvertisements- Sponsored content, promotionscalls_to_action- Subscribe, like, share promptspolitical_content- Political discussionsnegative_content- Drama, controversieskids_content- Children-focused contentself_promotion- Personal products, coursesrepetitive_content- Repeated explanationsfiller_speech- Um, uh, like, basicallytechnical_jargon- Complex technical detailspersonal_stories- Personal anecdotes
| Metric | Before (OpenAI) | After (Groq) | Improvement |
|---|---|---|---|
| Average Response Time | ~3-8 seconds | ~0.5-2 seconds | 75% faster |
| Context Window | 128K tokens | 128K tokens | Same capacity |
| Model Quality | GPT-4o-mini | Llama 4 Scout | Enhanced accuracy |
| Cost per 1M tokens | $0.15/$0.60 | $0.11/$0.34 | 43% cheaper |
- Short videos (<5 min): 0.5-1 seconds
- Medium videos (5-30 min): 1-2 seconds
- Long videos (30+ min): 2-4 seconds
- Cache hits: <0.1 seconds
# Test basic functionality
python test_groq.py
# Test specific video
curl -X GET "http://localhost:8000/process_video?video_id=dQw4w9WgXcQ"
# Test with preferences
curl -X POST "http://localhost:8000/process_video" \
-H "Content-Type: application/json" \
-d '{"video_id": "dQw4w9WgXcQ", "user_preferences": {"sensitivity": "high"}}'curl -X GET "http://localhost:8000/health"/health- Service health and cache status/api/stats- Detailed usage statistics- Cache management with automatic expiry
# Symptom: 401/403 errors
# Solution: Check GROQ_API_KEY environment variable
docker exec app env | grep GROQ_API_KEY# Check logs
docker-compose logs app
# Common issues:
# - Missing API key
# - Port 8000 already in use
# - Docker daemon not running# Check model status
curl -X GET "http://localhost:8000/api/stats"
# Possible causes:
# - Groq API rate limiting
# - Large video processing
# - Cold start after inactivity# Clear specific video cache
curl -X DELETE "http://localhost:8000/cache/VIDEO_ID"
# Check cache size
curl -X GET "http://localhost:8000/health"# Set production environment
export DEV_MODE=false
export GROQ_API_KEY=your_production_key
# Deploy
docker-compose up -d --build- Adjust cache expiry in
app.py(default: 24 hours) - Modify confidence thresholds for skip detection
- Customize prompts for specific content types
For high-traffic deployments:
# docker-compose.yml
services:
app:
deploy:
replicas: 3
# ... rest of config- Fork the repository
- Create a feature branch
- Make your changes
- Test with
python test_groq.py - Submit a pull request
This project is licensed under the MIT License - see the LICENSE file for details.
For questions or support, please open an issue on GitHub.
This project leverages Groq's ultra-fast inference and Meta's Llama 4 Scout model to deliver cutting-edge video content analysis. The AI components provide accurate identification and categorization of non-essential video segments for an optimized viewing experience.