Native Integration

Chat interface lives directly in Obsidian, matching its look and feel perfectly.

Instant Access

No switching between terminal and editor. Ask questions while writing.

Context Aware

Plugin knows which vault you're in, automatically using the right graph.

Easy Distribution

Share with others - they just enable the plugin and start chatting.

Frontend (Obsidian Plugin)

• Language: JavaScript
• UI: Custom chat interface
• Communication: REST API calls
• Files: main.js, styles.css, manifest.json

Backend (Python API)

• Framework: Flask
• Port: Auto-detected (5000+)
• Endpoints: /query, /health
• File: api_server.py

Auto Port Detection

def find_free_port(start_port=5000):
    for port in range(start_port, start_port + 10):
        try:
            sock = socket.socket()
            sock.bind(('', port))
            sock.close()
            return port
        except OSError:
            continue

Config Persistence

def save_port_config(port):
    config = {
        "port": port,
        "url": f"http://localhost:{port}"
    }
    with open(Path.home() / ".g-retriever-config.json", 'w') as f:
        json.dump(config, f)

Query Endpoint

@app.route('/query', methods=['POST'])
def query():
    data = request.get_json()
    question = data.get('question', '')
    result = retriever.query(question)
    return jsonify(result)

Health Check

@app.route('/health', methods=['GET'])
def health():
    return jsonify({
        'status': 'ok',
        'message': 'G-Retriever running'
    })

Factual Questions

• "What is gradient descent?"
• "How do I implement backpropagation?"
• "What are the main Python libraries I use?"

Connection Questions

• "How do neural networks relate to deep learning?"
• "What projects use React?"
• "Show connections between my ML notes"

Summary Questions

• "Summarize my notes on transformers"
• "What have I learned about Python?"
• "Overview of my project ideas"

Specific Lookups

• "Find notes mentioning PyTorch"
• "What did I write about attention mechanisms?"
• "Show me graph-related notes"

Custom Styling

Users can customize the chat appearance by editing styles.css:

/* Change chat colors */
.g-retriever-user {
background: #your-color;
}
/* Adjust sizes */
.g-retriever-input {
min-height: 100px;
}

Backend Configuration

Modify api_server.py to:

• Change default port range
• Adjust timeout values
• Add custom endpoints
• Enable debug logging

Performance Tuning

• Adjust k_retrieve parameter
• Change context length limit
• Modify temperature settings
• Cache frequently accessed nodes

Add Features

Extend the plugin with:

• Chat history export
• Favorite queries
• Multi-vault support
• Note insertion from chat

Architecture

Hybrid approach combines JavaScript frontend with Python backend for best of both worlds.

Integration

Using Obsidian's CSS variables ensures perfect theme compatibility.

⚡ User Experience

Auto-detection and clear status indicators make setup painless.

Distribution

Simple start scripts and clear README make sharing easy.