The database for AI
Graph, vector and analytics on the same records. Every write keeps a live schema of labels, fields and values, so agents query real fields instead of guessing.
Website • Docs • RushDB Cloud • Examples • Discussions
Bring data in the way you have it: nested JSON, CSV, SQL, Cypher or Mongo-style writes, or change streams from your databases. RushDB stores it as typed records linked into a graph and answers relationship queries, vector and full-text search, and aggregations from those same records, over HTTP, the PostgreSQL wire protocol, Bolt, gRPC or MCP. One Rust binary, Apache-2.0, or managed on RushDB Cloud.
-
Start a local database:
curl -fsSL https://rushdb.com/install.sh | sh mkdir my-app && cd my-app && ~/.rushdb/bin/rushdb up
Or
brew install rush-db/tap/rushdb, or Docker:docker run -p 127.0.0.1:3000:3000 -v rushdb:/data -e RUSHDB_TOKEN=change-me rushdb/rushdb:3
-
Add an SDK,
npm install @rushdb/javascript-sdkorpip install rushdb, and follow the quickstart. -
Working with an agent? Hand it rushdb.com/agent-setup, or add the MCP server.
-
Prefer managed? Sign up for RushDB Cloud.
| Area | Repositories |
|---|---|
| Core | rushdb (the database) · docs · examples |
| SDKs | javascript-sdk · rushdb-python |
| Agents | mcp-server · skills · agent-plugins · agent-memory-contract · rushdb-openclaw-memory · rushdb-hermes-memory |
| Frameworks | langchain-rushdb (LangChain, LangGraph) · llama-index-rushdb |
| Data in | connectors · debezium-server-rushdb · obsidian-plugin |
RushDB 2.x (legacy), the Neo4j-based stack, lives on the 2.x branch of rush-db/rushdb with security fixes until 2027-10-28.
