Vector Search Setup

Installation on DigitalOcean VPS

1. Install Ollama

curl -fsSL https://ollama.com/install.sh | sh

2. Pull the embedding model

ollama pull nomic-embed-text

This downloads ~274MB. The model will automatically start when needed.

3. Verify Ollama is running

curl http://localhost:11434/api/embeddings -d '{
  "model": "nomic-embed-text",
  "prompt": "test"
}'

Should return a JSON with an “embedding” array of 768 floats.

4. Restart your mu application

./mu

The app will now automatically:

• Generate embeddings for all indexed content (news, tickers, videos)
• Use semantic vector search for queries
• Fallback to keyword search if Ollama is unavailable

How it works

• Indexing: When news/tickers are indexed, embeddings are generated automatically
• Search: Queries are embedded and compared using cosine similarity
• Performance: ~50-100ms per embedding on 1-2 CPU cores
• Fallback: If Ollama is down, keyword search is used automatically

Testing

Try asking:

• “what’s the bitcoin price” → should find BTC ticker
• “ethereum value” → should find ETH ticker
  
• “crypto markets” → should find crypto-related news
• “digital gold” → should find Bitcoin content

Memory usage

• Ollama idle: ~100MB
• During embedding: +400MB temporarily
• Index with embeddings: ~4KB per entry