The database that thinks
One Datalog query language, every data model.
Built on LMDB, the fastest key-value ACID store.
Relational, graph, document, logical, vector, full-text in one unified DB.
Start embedded, move to cluster, never migrate data.
WAL and asynchronous transactions + state of the art query optimizer and rule engine.
Built-in embedding, text generation, and OCR, with full-text/vector/idoc retrieval and MCP tools.
Sharing ideas and helping each other with questions, examples, and issues.
If you enjoy Datalevin and it is helping you succeed, consider sponsoring the development and maintenance of this project.