rush-db/rushdb — RushDB is a graph + vector database and memory layer for AI agents.

Description and use

**Summary**
RushDB is an open‑source graph‑plus‑vector database built on Neo4j that lets AI agents store arbitrary JSON and retrieve typed, searchable, relationship‑aware records without needing a predefined schema or migrations. It provides a standard protocol for connecting AI assistants to tools and data, making it easy to expose Model Context Protocol (MCP) services or build custom integrations. With active maintenance, a growing community, and a TypeScript SDK/CLI, it is ready for serious pilot projects.

**Value**

• **Schema‑free, relationship‑aware storage**: Developers can push any JSON payload and later query it with rich graph semantics and vector similarity, eliminating the overhead of schema design and migration cycles.
• **Unified protocol for AI tooling**: By exposing a consistent API/SDK, RushDB becomes a common “data glue” that lets diverse AI agents (RAG pipelines, autonomous assistants, tool‑calling bots) interact with real‑world data and external services through a single interface.
• **Built on Neo4j**: Leverages Neo4j’s mature graph engine for reliability, scalability, and ACID guarantees while adding vector search capabilities needed for modern LLM‑driven workflows.

**Practical adoption path**

1. **Prototype** – Install the TypeScript SDK or use the provided CLI to spin up a local RushDB instance, push sample JSON documents, and test graph + vector queries.
2. **Integrate** – Replace ad‑hoc data stores in your AI agent’s tool‑calling layer with RushDB calls (e.g., push, search, traverse). The SDK abstracts the underlying Neo4j details, so integration is a few lines of code.
3. **Deploy** – Deploy RushDB in a containerized environment (Docker/K8s) using the official Helm chart or Docker image; configure authentication and TLS as per your security policies.
4. **Scale** – Leverage Neo4j’s clustering features for high‑availability and horizontal scaling as query volume grows; the vector index can be tuned independently for performance.

**Production readiness**

• **Activity & community**: 229 stars, recent commits (as of 2026‑06‑23), 20+ topics, and active issue response indicate a healthy open‑source project.
• **Maturity**: The combination of Neo4j’s proven graph engine and a TypeScript front‑end gives strong reliability and type safety.
• **Ecosystem fit**: Clear API/SDK/CLI surface, support for MCP, and straightforward Docker/Helm deployment make it easy to evaluate and roll out in pilot environments.
• **Risks**: Final due diligence is needed on licensing (check the exact OSS license), security posture (review CVEs for Neo4j and the RushDB codebase), and maintainership continuity, but no major red flags have been identified.

Overall, RushDB is a production‑ready OSS candidate for teams looking to give AI agents a unified, schema‑free graph‑vector store and a standard integration point for real‑world tools and data.

Problem

rush-db/rushdb helps connect AI assistants to real tools and data through a standard protocol.

Outcome

High for an OSS candidate: recent activity, adoption, and ecosystem signals are strong enough for a serious pilot.

Workflow

• connect AI agents to tools
• ship Model Context Protocol servers
• standardize integrations

Evidence

RushDB is a graph + vector database and memory layer for AI agents. Push any JSON, get typed, searchable, relationship-aware records back — no schema, no migrations. Built on Neo4j.

Source