The mcp-memory-libsql server is a high-performance, persistent memory system with vector search and knowledge graph capabilities powered by libSQL.
Entity Management: Create, update, and delete entities with optional embeddings for similarity search
Relation Management: Create and delete relations between entities
Search: Perform text or vector similarity searches on entities and relations
Graph Querying: Retrieve and explore the knowledge graph of entities and their relations
Persistence: Store data using local SQLite or remote libSQL databases like Turso
Authentication: Secure token-based authentication for remote databases
Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type @ followed by the MCP server name and your instructions, e.g., "@mcp-memory-libsql search for entities related to project planning and deadlines"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
A high-performance, persistent memory system for the Model Context Protocol (MCP) powered by libSQL with optimized text search for LLM context efficiency.
🚀 High-performance text search with relevance ranking
💾 Persistent storage of entities and relations
🔍 Flexible text search with fuzzy matching
🎯 Context-optimized for LLM efficiency
🔄 Knowledge graph management
🌐 Compatible with local and remote libSQL databases
🔒 Secure token-based authentication for remote databases
Related MCP server: dragon-brain
This server is designed to be used as part of an MCP configuration. Here are examples for different environments:
Add this to your Cline MCP settings:
{
"mcpServers": {
"mcp-memory-libsql": {
"command": "npx",
"args": ["-y", "mcp-memory-libsql"],
"env": {
"LIBSQL_URL": "file:/path/to/your/database.db"
}
}
}
}For a detailed guide on setting up this server with Claude Desktop in WSL, see Getting MCP Server Working with Claude Desktop in WSL.
Add this to your Claude Desktop configuration for WSL environments:
{
"mcpServers": {
"mcp-memory-libsql": {
"command": "wsl.exe",
"args": [
"bash",
"-c",
"source ~/.nvm/nvm.sh && LIBSQL_URL=file:/path/to/database.db /home/username/.nvm/versions/node/v20.12.1/bin/npx mcp-memory-libsql"
]
}
}
}The server supports both local SQLite and remote libSQL databases through the LIBSQL_URL environment variable:
For local SQLite databases:
{
"env": {
"LIBSQL_URL": "file:/path/to/database.db"
}
}For remote libSQL databases (e.g., Turso):
{
"env": {
"LIBSQL_URL": "libsql://your-database.turso.io",
"LIBSQL_AUTH_TOKEN": "your-auth-token"
}
}Note: When using WSL, ensure the database path uses the Linux
filesystem format (e.g., /home/username/...) rather than Windows
format.
By default, if no URL is provided, it will use file:/memory-tool.db
in the current directory.
The server implements the standard MCP memory interface with optimized text search:
Entity Management
Create/Update entities with observations
Delete entities
Search entities by text with relevance ranking
Explore entity relationships
Relation Management
Create relations between entities
Delete relations
Query related entities
The server uses a libSQL database with the following schema:
Entities table: Stores entity information with timestamps
Observations table: Stores entity observations
Relations table: Stores relationships between entities
Text search with relevance ranking (name > type > observation)
Due to npm 2FA requirements, publishing needs to be done manually:
Create a changeset (documents your changes):
pnpm changesetVersion the package (updates version and CHANGELOG):
pnpm changeset versionPublish to npm (will prompt for 2FA code):
pnpm releaseContributions are welcome! Please read our contributing guidelines before submitting pull requests.
MIT License - see the LICENSE file for details.
Built on the Model Context Protocol
Powered by libSQL
Create new entities with observations and optional embeddings
Inputs: entities (required).
Create relations between entities
Inputs: relations (required).
Delete an entity and all its associated data (observations and relations)
Inputs: name (required).
Delete a specific relation between entities
Inputs: source (required), target (required), type (required).
Get recent entities and their relations
Inputs: no named parameters.
Search for entities and their relations using text or vector similarity
Inputs: query (required).
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
Scored across 6 tools
Every tool has a clearly distinct purpose: create_entities and create_relations handle creation of different data types, delete_entity and delete_relation handle deletion of different data types, read_graph retrieves recent data, and search_nodes performs searches. There is no overlap or ambiguity between these functions.
All tools follow a consistent verb_noun pattern using snake_case: create_entities, create_relations, delete_entity, delete_relation, read_graph, and search_nodes. The naming is predictable and uniform throughout the set.
With 6 tools, this server is well-scoped for a memory/graph database system. Each tool serves a clear and necessary function (create, delete, read, search) without being too sparse or bloated, fitting typical expectations for such a domain.
The toolset covers core CRUD operations for entities and relations (create, delete, read, search), but lacks explicit update tools for modifying existing entities or relations. Agents can work around this by deleting and recreating, but it's a minor gap in lifecycle coverage.
Cloud-hosted MCP server for durable AI memory
Capability registry for the agentic economy. Semantic search over verified MCP server listings.
Shared cross-LLM long-term memory over MCP: semantic recall, sessions, and media (pgvector).
