AvocadoDB

A deterministic context database that solves the problem of inconsistent retrieval in RAG pipelines by ensuring the same query returns the same context every time. It targets developers and engineering teams building AI applications, particularly those using RAG, and is positioned as an open-source, local-first alternative to probabilistic vector databases, offering bit-exact results, full traceability, and a single binary with zero dependencies.

Key features

• Bit-exact deterministic results
• Hybrid search (HNSW + BM25)
• Local-first, no data leaks
• Git-like version control
• Sub-50ms query latency
• 10k/sec indexing speed
• Single binary, zero dependencies
• Seamless integration with LangChain, LlamaIndex