Skip to content

RecursionAI/FlowDB

Repository files navigation

FlowDB

FlowDB is a JSON-based database stack that is AI-Native.

It's core objective is to be fast, simple, and n8n + AI optimized.


Core Features

  • JSON (No-SQL)
  • API controlled for n8n-usage and database safety
  • API Key protected
  • Dockerized for ease of use
  • MCP server for Agentic use
  • Vectorization built in (BYOK)
  • Client CLI and PyPi package
  • Speed: Built on LMDB (Lightning Memory-Mapped Database) and USearch (HNSW Vector Search).

Guidelines

FlowDB is intentionally built to be fast, simple, and AI-Native. It is NOT ideal for compliance applications such as FinTech because it lacks SQLs ACID safety measures.

FlowDB is ideal for AI applications, automation, speed, and n8n workflows.


Getting Started

Full guide to install both the Client SDK and configure the dockerized server.

How to use the client SDK to interact with the database in your code and utilyze Pydantic models

Comprehensive guide for implementing FlowDB in your n8n workflows for and open-source persistent memory adn VectorDB solution.

URL schemas and API guidelines.

How the MCP server works and how to set it up so Claude Desktop can use it and interact with your db.


About

Fast, Simple, Agentic Database

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors