rushdb

The database for AI: graph, vector and analytics on the same records, with a live schema of labels, fields and values that agents read before they query. One Ru

What is rushdb?

The database for AI: graph, vector and analytics on the same records, with a live schema of labels, fields and values that agents read before they query. One Rust binary, Apache-2.0.

Overview of rushdb

What it is

RushDB is a graph + vector database and memory layer for AI agents. It accepts any JSON input and returns typed, searchable, relationship-aware records without requiring a predefined schema or migrations. The database is built on Neo4j and replaces the need for separate Redis, vector store, and graph database systems, along with the glue code to keep them synchronized.

Key features

  • Managed embeddings: Index any string property once; every write is auto-embedded server-side.
  • Graph + vector in one query: Semantic similarity and relationship traversal compose in a single call.
  • Zero schema: Push any JSON shape; RushDB infers types, creates properties, and links records automatically.
  • Inferred schema: Properties become first-class nodes; types, labels, and relationships are discovered on write, building a queryable schema as data arrives.
  • ACID transactions: Concurrent agents do not corrupt shared memory, thanks to Neo4j under the hood.
  • SDKs for JavaScript and Python, plus an MCP server for integration with Claude, Cursor, and other MCP clients.

Use cases

RushDB is designed for AI agents that need persistent, structured memory. It can store and recall agent actions by meaning using combined graph filters and semantic search. It supports importing nested JSON and CSV data, traversing relationships arbitrarily deep in a single query, and performing analytical queries with aggregations and grouping. The MCP server allows agents to create records, search by meaning, traverse relationships, and introspect the schema via natural language.

Getting started

Two paths are available: Cloud (managed, free tier) and self-host (Docker + your own Neo4j instance). Install the SDK: bash npm install @rushdb/javascript-sdk # or pip install rushdb For full quick-start instructions, including API key setup and example code, see the repository README.

License and pricing

The project is listed as open source (pricing_type: open_source). The license is not explicitly stated in the repository metadata or README.

Based on the project's public README and repository metadata.

Add this badge to your README

Featured on YunoDB Badge styles & help

Maintain this project? Show that it is featured on YunoDB.

Markdown[![Featured on YunoDB](https://yunodb.com/badge/rush-dbrushdb.svg)](https://yunodb.com/tools/rush-dbrushdb?utm_source=github&utm_medium=badge)
HTML<a href="https://yunodb.com/tools/rush-dbrushdb?utm_source=github&utm_medium=badge"><img src="https://yunodb.com/badge/rush-dbrushdb.svg" alt="Featured on YunoDB" height="20"></a>