Show HN: ChimeraDB – Vector search, graph queries and SQL

I built ChimeraDB to stop juggling three separate databases when building LLM apps. It combines vector embeddings, property graphs, and SQL analytics in a single DuckDB file.

    from chimeradb import KnowledgeGraph
    kg = KnowledgeGraph("my.db")
    
    # Semantic search - find by meaning
    results = kg.search("who works on language models?")
    
    # Graph traversal - follow relationships  
    employees = kg.traverse("acme", direction="incoming")
    
    # SQL analytics - aggregate data
    stats = kg.query("SELECT company, COUNT(*) FROM nodes...")

Why it's useful:

- RAG needs semantic search + relationship context

- AI agents need graph traversal + analytical queries

- Combine all three in a single SQL query

Zero infrastructure: One file, runs anywhere, 10-100x faster than SQLite for analytics.

Built on DuckDB + duckpgq + vss extensions. MIT licensed.

    pip install chimeradb

GitHub: https://github.com/codimusmaximus/chimeradb