Repolode
Submit
openmemind

openmemind/memind

Java-native long-term memory and context engine for AI agents, turning conversations, documents and tool calls into structured, retrievable memory.

View on GitHubWebsite
Stars
901
Forks
96
Open issues
13
Last commit
2 months ago

What is it?

Memind is a self-evolving memory layer for AI applications written in Java. It captures raw context from conversations, documents, images, audio, tool calls and agent timelines, then extracts user memory, reusable agent experience and higher-level insights from them.

Stored memory is organized into Memory Items, Memory Graphs that link entities and temporal or causal relations, Memory Threads that group related work, and Insight Trees that distill raw memories from leaf to root. Retrieval combines vector search, BM25 keyword search and graph expansion, with an optional Deep Retrieval mode that adds query expansion and reranking.

Memind can be used as a standalone server with REST APIs, an HTTP MCP endpoint and SDKs in several languages, or embedded in a Java or Spring Boot application. First-party integrations are provided for Claude Code, Codex, OpenClaw and Hermes.

Key features

  • 01Extracts structured user memory and reusable agent experience from conversations and tool calls.
  • 02Builds Insight Trees, memory graphs and memory threads that link related context over time.
  • 03Hybrid retrieval across vector search, BM25 keyword search and graph expansion.
  • 04Exposes REST APIs, an HTTP MCP server and official clients for TypeScript, Python, Java, Go and Rust.
  • 05Runs embedded in plain Java or Spring Boot, with SQLite, MySQL and PostgreSQL storage plugins.

Use cases

  • Giving a coding agent persistent project context and memory of resolved tasks.
  • Adding long-term user profiles and preferences to a chatbot or companion application.
  • Letting a Java backend store and query memory for a workflow or support agent.
  • Connecting an MCP-capable agent to a shared, scoped memory service.

Quick start

  1. 1

    Copy the example environment file

    cp .env.example .env
  2. 2

    Edit .env and set your model provider API key and model names

  3. 3

    Start memind-server and the admin UI with Docker Compose

    docker compose up -d --build
  4. 4

    Check that the server is running

    curl http://localhost:8366/open/v1/health

Things to know

  • Running it requires a valid API key from a model provider for chat and embedding models.
  • The admin UI has no built-in authentication, so it should not be exposed to public networks.

Topics

#ai-agent#ai-memory#java#spring-ai#context-engineering#vector-search#mcp#rag#ai#ai-agents

FAQ

What is memind?
Java-native long-term memory and context engine for AI agents, turning conversations, documents and tool calls into structured, retrievable memory. Memind is a self-evolving memory layer for AI applications written in Java. It captures raw context from conversations, documents, images, audio, tool calls and agent timelines, then extracts user memory, reusable agent experience and higher-level insights from them.
Is memind free to use? What license does it have?
memind is open source under the Apache-2.0 license. Check the license terms before using it in commercial products.
Is memind still maintained?
Updated within 6 months. The last commit to memind was on Aug 13, 2026.
What can I use memind for?
Giving a coding agent persistent project context and memory of resolved tasks. Adding long-term user profiles and preferences to a chatbot or companion application. Letting a Java backend store and query memory for a workflow or support agent. Connecting an MCP-capable agent to a shared, scoped memory service.
How do I install memind?
Copy the example environment file: cp .env.example .env Edit .env and set your model provider API key and model names Start memind-server and the admin UI with Docker Compose: docker compose up -d --build Check that the server is running: curl http://localhost:8366/open/v1/health
Where can I get memind?
The source code and installation instructions are on GitHub: https://github.com/openmemind/memind

Summary generated by AI from the project's README. GitHub stats refresh daily.

Related projects