Devmind
DevMind MCP provides **persistent memory capabilities** for AI assistants through the Model Context Protocol (MCP). It enables AI to remember context across conversations, automatically track development activities, and retrieve relevant information intelligently.
claude mcp add --transport stdio jochenyang-devmind-mcp npx -y devmind-mcp \ --env DM_MCP_STORAGE_PATH="Path to local SQLite storage or directory for data (optional)"
How to use
DevMind MCP is a persistent memory and context management server that implements the Model Context Protocol (MCP) to empower AI assistants with memory across conversations. It combines semantic search, keyword search, and context-aware memory management to provide a private, local storage solution built on SQLite. The server exposes a toolkit of 15 MCP-powered features, including session and context management, project indexing, codebase analysis, and visualization tools, all designed to integrate with MCP clients like Claude Code or Cursor. To start using it, run the MCP server via a package manager (for example with npx) and connect your MCP client to the server to begin recording contexts, querying memories, and leveraging the codebase indexing for efficient retrieval. Once running, you can perform memory-related operations, search across contexts semantically, index your codebase for quick discovery, and visualize relationships between memories and projects through the provided tools.
How to install
Prerequisites:
- Node.js version 20.0.0 or newer
- Access to npm (comes with Node.js)
Installation steps:
-
Install the MCP server package (globally) so you can run it from the command line:
npm install -g devmind-mcp
-
Verify installation by checking the version (optional):
npm -g ls devmind-mcp
-
Start the MCP server using npx (no global install required):
npx -y devmind-mcp
-
If you prefer running directly, you can also install locally in a project:
mkdir my-devmind-project cd my-devmind-project npm init -y npm i --save-dev devmind-mcp npx -y devmind-mcp
Prerequisites recap:
- Node.js ≥ 20.x
- npm or yarn as your package manager
- Basic familiarity with MCP clients to connect to the running server
Additional notes
Tips and common considerations:
- The server stores data locally using SQLite for privacy; ensure the storage path has adequate disk space.
- You can customize storage paths and certain behaviors via environment variables described in the mcp_config (e.g., DM_MCP_STORAGE_PATH).
- If you run into port or network issues, verify that MCP client connections are allowed to reach the running process and that any firewall rules permit the IPC or HTTP/S endpoints used by your MCP client.
- When upgrading the DevMind MCP package, review release notes for breaking changes related to the 15 MCP tools, auto-memory tiers, or storage behavior.
- For development, you can rebuild or recompile TypeScript sources if you clone the repo locally, and then run the updated server binary via the same command pattern.
- If you need to customize how memory is recorded (tier behavior, force_remember flags, etc.), refer to the project’s documentation on Type-Based Auto-Memory and the smart memory components.
Related MCP Servers
mcp-svelte-docs
🔍 MCP server that lets you search and access Svelte documentation with built-in caching
augments
Comprehensive MCP server providing real-time framework documentation access for Claude Code with intelligent caching, multi-source integration, and context-aware assistance.
toolkit
A Model Context Protocol server providing LLM Agents with system utilities and tools, including IP geolocation, network diagnostics, system monitoring, cryptographic operations, and QR code generation.
mini_claude
Give Claude Code persistent memory across sessions. Track habits, log mistakes, prevent death spirals. Runs locally with Ollama.
mcp-github-integration
A TypeScript package for interacting with the GitHub API through an MCP (Model Context Protocol) server integration.
warp-sql
🗄️ Model Context Protocol (MCP) server for SQL Server integration with Warp terminal. Execute queries, explore schemas, export data, and analyze performance with natural language commands.