An intelligent desktop news aggregation application that uses local AI to filter, categorize, and analyze news content with political bias detection.

🤖 AI-Powered Content Analysis

• Automatic article categorization (technology, politics, business, science, health, sports, entertainment, world)
• Sentiment analysis (positive, negative, neutral)
• Importance scoring (high, medium, low)
• Topic extraction and tagging

🎯 Political Bias Detection

• Bias scoring on a -1.0 to 1.0 scale (left to right)
• Confidence levels for bias assessments
• Detailed reasoning explaining AI's bias determination
• Visual bias indicators with color-coded scales

📊 Smart Content Discovery

• RSS feed parsing and management
• Relevance scoring based on user preferences
• AI-generated summaries on demand
• User interaction tracking for personalization

🔒 Privacy-Focused

• Local AI processing with LM Studio
• No external API calls for content analysis
• Local SQLite database storage

• Frontend: HTML/CSS/JavaScript with responsive design
• Backend: Python FastAPI server
• Database: SQLite with embedding storage
• AI: LM Studio (local LLM) for content analysis
• Future: Tauri desktop app framework

• Python 3.8+
• LM Studio running locally
• WSL2 (if on Windows)

1. Clone the repository:

git clone <repository-url>
cd news_feed

2. Install Python dependencies:

pip install --user fastapi uvicorn feedparser requests openai==0.28.1 numpy

3. Set up LM Studio:

   • Install and run LM Studio
   • Load a model (e.g., Mistral 7B Instruct)
   • Enable "Network Access" and "Enable CORS" in settings
   • Ensure it's running on port 1234

4. Initialize the database:

python3 backend/main.py

1. Start the backend server:

export PATH="/home/jon/.local/bin:$PATH"
python3 backend/main.py

2. Open the frontend:

# Open src/index.html in your browser
# Or serve it locally:
cd src
python3 -m http.server 3000

3. Access the application at http://localhost:3000

• Refresh Feeds: Click "🔄 Refresh Feeds" to fetch new articles
• Filter by Category: Use the dropdown to filter articles by category
• AI Summary: Click "🤖 AI Summary" on any article for an AI-generated summary
• Bias Analysis: Click "🎯 Bias Analysis" on political articles to see detailed bias assessment
• Read Full Article: Click "Read Full" to open the original article
• Rate Articles: Use 👍/👎 buttons to help train the relevance algorithm

Default feeds are automatically added on first run. To add custom feeds:

curl -X POST "http://localhost:8000/feeds" \
  -H "Content-Type: application/json" \
  -d '{"feed_url": "https://example.com/rss"}'

For WSL users, the application automatically detects the Windows host IP. If you're running everything locally, it will use localhost:1234.

Manual configuration in backend/ai/lm_studio_client.py:

self.base_url = "http://your-lm-studio-host:1234"

• GET / - API information and health status
• GET /articles - Get filtered articles with pagination
• GET /feeds - Get all active RSS feeds
• POST /feeds - Add a new RSS feed
• POST /refresh - Refresh all feeds and process with AI
• POST /articles/{id}/summary - Generate AI summary for article
• GET /articles/{id}/bias-analysis - Get detailed bias analysis
• POST /articles/{id}/interact - Record user interaction
• GET /health - Health check and LM Studio status

• Basic metadata (title, URL, content, author, published date)
• AI analysis results (category, sentiment, importance, topics)
• Political bias data (bias score, confidence, reasoning)
• User interaction data (relevance score)
• Vector embeddings for similarity matching

• RSS feed metadata and management
• Active/inactive status and last fetch times

• Tracks user behavior (views, likes, dislikes, read time)
• Used for personalization and relevance scoring

news_feed/
├── backend/
│   ├── ai/
│   │   └── lm_studio_client.py    # LM Studio integration
│   ├── content/
│   │   └── feed_parser.py         # RSS parsing
│   ├── database/
│   │   └── models.py              # Database operations
│   └── main.py                    # FastAPI application
├── src/
│   └── index.html                 # Frontend application
├── CLAUDE.md                      # Project documentation
└── README.md                      # This file

Currently no automated tests. Manual testing via the web interface.

1. Fork the repository
2. Create a feature branch
3. Make your changes
4. Test thoroughly
5. Submit a pull request

"Backend Offline" status:

• Check if the Python server is running on port 8000
• Verify no port conflicts: lsof -i:8000

"AI Offline" status:

• Ensure LM Studio is running and has a model loaded
• Check LM Studio settings: Enable "Network Access" and "Enable CORS"
• For WSL: Verify Windows firewall allows port 1234

Database errors:

• Run the migration script: python3 add_bias_reasoning_column.py
• Delete news_feed.db to reset (will lose data)

"Loading articles..." stuck:

• Restart the backend server
• Check browser console for JavaScript errors
• Verify API endpoints are responding: curl http://localhost:8000/health

Bias analysis not working:

• Ensure articles have been processed with AI (refresh feeds)
• Check that LM Studio is available and responding

MIT License - see LICENSE file for details.

• LM Studio for local AI processing
• FastAPI for the backend framework
• RSS feed providers for content sources