selfhost.directory

The index of self-hostable projects — ranked, categorized and served with live updates.
Own your data and stop giving it to big tech.

rushdb logo

Self-host rushdb

@rushdb/mcp-server@2.13.0

Unified graph and vector database for AI agent memory

BetaOpen Core
Updated9 hours ago
LanguageTypescript
Setup difficultyAdvanced
Stars
325
Port
3000 (HTTP API)
Requires
APOC plugin for Neo4j, Neo4j
Docker
rushdb/platform
Install with Docker

What is rushdb?

RushDB is a context layer for AI agents, applications, and analytics that unifies graph relationships, vector embeddings, and live schema inference into a single queryable model. You push any JSON and RushDB automatically creates typed records, inferred relationships, and server-side embeddings — no schema design or migrations needed. It exposes a unified query API for graph traversal, semantic search, aggregations, and schema introspection. The self-hosted edition runs as a Docker container (rushdb/platform) connected to your own Neo4j 2026.01.4+ instance with the APOC plugin. It includes an MCP server for Claude/Cursor integration, TypeScript and Python SDKs, and a CLI for user management. The platform core and dashboard are licensed under Elastic License 2.0, while SDKs and documentation use Apache 2.0. It replaces the typical stack of Redis + vector store + graph DB + glue code with one API.

Should you self-host rushdb?

rushdb scores 48/100 on selfhost.directory (how scoring works) and ranks #5 of 5 in Vector Database. It has 325 GitHub stars.

It shipped 32 stable releases in the past year. The latest, @rushdb/mcp-server@2.13.0, came out on 31 August 2026.

Setup is advanced: plan for several services and some ongoing maintenance. An official Docker image is available. It is written mostly in Typescript.

Its closest alternative on selfhost.directory is qdrant (82/100, 35 298 stars). It scores higher overall, so compare the two before you pick. See rushdb vs qdrant side by side.

Features

  • ACID Transactions
  • Agent Skills
  • Analytical queries (aggregations)
  • CLI user management
  • CSV import with type inference
  • Graph Traversal
  • Graph + vector search in one query
  • Inferred live schema
  • Managed Embeddings
  • MCP Server
  • Nested JSON import
  • Schema introspection
  • Self-hosted with Docker
  • Unified Query API
  • Zero schema / schema-free

rushdb README

How to install rushdb

Install rushdb with Docker

Environment variables
NEO4J_URL, NEO4J_USERNAME, NEO4J_PASSWORD

docker run -p 3000:3000
--name rushdb
-e NEO4J_URL='neo4j+s://your-instance.databases.neo4j.io'
-e NEO4J_USERNAME='neo4j'
-e NEO4J_PASSWORD='password'
rushdb/platform

Requires external Neo4j instance

How to configure rushdb

  • NEO4J_PASSWORDNeo4j password

    Neo4j password

  • NEO4J_URLNeo4j connection URL

    Neo4j connection URL

    values: neo4j+s://..., bolt://...

  • NEO4J_USERNAMEneo4jNeo4j username

    Neo4j username

  • RUSHDB_AES_256_ENCRYPTION_KEY32-char key for API token encryption

    32-char key for API token encryption

    values: 32-character string

  • RUSHDB_BASE_URLPublic/base API URL for synx assignments

    Public/base API URL for synx assignments

  • RUSHDB_LOGINadminAdmin login

    Admin login

  • RUSHDB_PASSWORDpasswordAdmin password

    Admin password

  • RUSHDB_PORT3000HTTP port

    HTTP port

  • RUSHDB_SYNX_CONTROL_TOKENInternal token for managed synx workers

    Internal token for managed synx workers

  • RUSHDB_SYNX_DESTINATION_API_KEYInternal write key for synx destination writes

    Internal write key for synx destination writes

rushdb releases & version history

FAQ

Is RushDB a database or a layer over my existing database?

Both deployment patterns are supported. Use RushDB as the primary backend when it can own connected, evolving operational records. Use it as a shared layer when records still originate in PostgreSQL, MongoDB, object storage, or other systems of record.

More in Graph Database

Rate this project

Click to rate — no account needed.

0 comments

Posting as guest
Loading comments…