Show HN: LatticeDB – Like SQLite but for graph databases

7.7 relevance

New graph database tool with SQLite-like simplicity, directly relevant to data engineering and open source.

AI/ML github.com

Answer the board on AI readiness

Arrive with evidence, not opinion. See how your foundation benchmarks against your peers, get a clear read, and leave with a ranked place to start.

Why It Matters

For platform engineers building agent memory, Graph RAG, or local knowledge tools, LatticeDB eliminates the operational overhead of running separate graph, vector, and search databases while keeping data portable in a single file.

Key Takeaways

Evaluate LatticeDB for local-first AI workloads that need combined graph traversal, vector similarity, and full-text search without managing a server.

Embedded single-file knowledge graph database with vector search and full-text search for AI/RAG apps - jeffhajewski/latticedb

Summary

LatticeDB is an embedded, single-file property-graph database that unifies graph traversal, HNSW vector similarity search, and BM25 full-text indexing in one query layer. It achieves 0.13 μs node lookups and 0.83 ms vector search at 1M vectors with 100% recall, using a WAL-backed single-writer model for local-first applications. The engine supports Python, TypeScript, and Go bindings, with built-in durable event streams and a graph changefeed sharing the same transaction path as writes.

Author

jeffhajewski