MindOS

Name: MindOS
Author: GeminiLight

@GeminiLight

About MindOS

Local-first knowledge base MCP server. Let AI agents (Claude, Cursor, Gemini CLI, etc.) read and write your Markdown notes via MCP protocol. Supports Streamable HTTP and stdio transports.

#knowledge-base #mcp-server #ai-agent #markdown #local-first

Basic information

Config

Add this server to your MCP-compatible client using the configuration below.

{
  "mcpServers": {
    "mindos": {
      "type": "http",
      "url": "http://localhost:3000/mcp",
      "headers": {
        "Authorization": "Bearer YOUR_AUTH_TOKEN"
      }
    }
  }
}

Tools

No tools detected

We auto-extract tools from the README. The maintainer can list them under a ## Tools heading to populate this section.

Overview

What is MindOS?

MindOS is a local-first knowledge base that lets AI agents read and write Markdown notes via the Model Context Protocol (MCP). It provides persistent memory for agents, stored on the user’s own machine in plain Markdown files.

How to use MindOS?

Install the package globally with npm install -g @geminilight/mindos, then run mindos onboard for interactive setup and mindos start to launch the app (port 3000) and MCP server (port 8787). Add the MCP server URL http://localhost:8787/mcp with a bearer token to your agent’s MCP configuration.

Key features of MindOS

Local-first storage using plain Markdown files
20 MCP tools for reading and writing notes
Full-text search across the knowledge base
Git history tracking for each file
Backlink discovery between files
Supported by Claude Code, Cursor, Gemini CLI, Windsurf, and Cline

Use cases of MindOS

Give an AI agent a persistent, editable knowledge base on your local machine
Allow agents to search and retrieve information from your notes
Enable agents to create, update, and delete Markdown and CSV files
Track changes and history using built-in git integration
Provide startup context for agents via the bootstrap tool

FAQ from MindOS

—

Comments

More AI & Agents MCP servers

Gemini MCP Server

aliargun

MCP server implementation for Google's Gemini API

259

欢迎来到智言平台

Shy2593666979

AgentChat 是一个基于 LLM 的智能体交流平台，内置默认 Agent 并支持用户自定义 Agent。通过多轮对话和任务协作，Agent 可以理解并协助完成复杂任务。项目集成 LangChain、Function Call、MCP 协议、RAG、Memory、HITL、Skill、Milvus 和 ElasticSearch 等技术，实现高效的知识检索与工具调用，使用 FastAPI 构建高性能后端服务。

797

Model Context Protocol for Unreal Engine

chongdashu

Enable AI assistant clients like Cursor, Windsurf and Claude Desktop to control Unreal Engine through natural language using the Model Context Protocol (MCP).

2.0K