mcp-memory-libsql
đź§ High-performance persistent memory system for Model Context Protocol (MCP) powered by libSQL. Features vector search, semantic knowledge storage, and efficient relationship management - perfect for AI agents and knowledge graph applications.
claude mcp add --transport stdio spences10-mcp-memory-libsql npx -y mcp-memory-libsql \ --env LIBSQL_URL="file:/path/to/database.db" \ --env LIBSQL_AUTH_TOKEN="<your-auth-token-if-using-remote-libsql>"
How to use
The mcp-memory-libsql server provides a high-performance, persistent memory store backed by libSQL with optimized text search for efficient retrieval of entities, observations, and relationships. It implements the standard MCP memory interface, enabling you to create, update, delete, and search for entities and their observations, as well as manage relations between entities. The server supports both local SQLite databases and remote libSQL endpoints via the LIBSQL_URL environment variable, giving you flexibility for on-disk or remote, scalable storage. To use it, configure an MCP environment to reference this server entry (for example via a configuration like the included JSON snippets) and provide the LIBSQL_URL (and LIBSQL_AUTH_TOKEN for remote databases) as needed. The server exposes capabilities for entity management (create/update, delete, text search with relevance ranking), relation management (create/delete relations, query related entities), and graph exploration, all optimized for LLM-context efficiency.
How to install
Prerequisites:
- Node.js (recommended LTS) and npm or pnpm installed on your system
- Access to the internet to fetch npm packages
Installation and setup steps:
-
Ensure Node.js is installed. Verify: node -v npm -v
-
Install and run the MCP server using npx (as shown in the configuration example): npx -y mcp-memory-libsql
This will install the package and start the server with default settings. If you prefer to pass environment variables directly, you can prepend them, e.g.: LIBSQL_URL=file:/path/to/database.db npx -y mcp-memory-libsql
-
Alternatively, if you maintain a dedicated configuration file (as per the MCP standard), place the following in your MCP config under mcpServers.mcp-memory-libsql: { "command": "npx", "args": ["-y", "mcp-memory-libsql"], "env": { "LIBSQL_URL": "file:/path/to/database.db", "LIBSQL_AUTH_TOKEN": "<token>" } }
-
If you’re using WSL or a remote environment, ensure paths and shell commands are compatible with your setup (see README’s Claude Desktop/WSL configuration as guidance).
-
Start your MCP configuration and verify the server comes up and responds to MCP memory interface calls.
Additional notes
Tips and notes:
- The LIBSQL_URL environment variable controls the database location. For local development, you can use file:/path/to/database.db. For remote libSQL services (e.g., Turso), use libsql://<host> and provide LIBSQL_AUTH_TOKEN.
- If no LIBSQL_URL is provided, the server defaults to an in-memory database at ./memory-tool.db.
- When using WSL, prefer Linux-style paths for database files (e.g., /home/username/...).
- This server exposes typical MCP memory operations: create/update/delete/search entities and observations, and create/delete/explore relations with relevance-ranked text search.
- If you’re integrating with Claude Desktop or other tools, you can mirror the configuration approaches shown for WSL in the README to ensure compatibility.
- For publishing and versioning, this project follows npm-based publishing with 2FA (pnpm changeset, etc.).
Related MCP Servers
augments
Comprehensive MCP server providing real-time framework documentation access for Claude Code with intelligent caching, multi-source integration, and context-aware assistance.
mcp-n8n-builder
🪄 MCP server for programmatic creation and management of n8n workflows. Enables AI assistants to build, modify, and manage workflows without direct user intervention through a comprehensive set of tools and resources for interacting with n8n's REST API.
vikunja
Model Context Protocol server for Vikunja task management. Enables AI assistants to interact with Vikunja instances via MCP.
mcp -text-editor
An open source implementation of the Claude built-in text editor tool
qdrant
MCP server for semantic search using local Qdrant vector database and OpenAI embeddings
grok-faf
First MCP server for Grok | FAST⚡️AF • URL-based AI context • Vercel-deployed