Typesense MCP Server

A Model Context Protocol (MCP) server implementation that connects AI models to Typesense search capabilities, enabling LLMs to discover, search, and analyze data within Typesense collections.

Features

Resources

• Lists and provides access to collections via typesense:// URIs, including name, description, and document count.
• Offers JSON mime type for schema access.

Tools

• typesense_query: Searches documents with filtering, returning matching documents with relevance scores.
• typesense_get_document: Retrieves specific documents by ID.
• typesense_collection_stats: Provides collection metadata, document count, and schema information.

Prompts

• analyze_collection: Analyzes collection structure and contents, providing insights about schema, data types, and statistics.
• search_suggestions: Recommends search strategies based on collection schema.

Installation

Install via npm (globally or locally) or using mcp-get.

Development

Install dependencies, build the server, or use auto-rebuild during development.

Installation for Development

Configuration instructions are provided for Claude Desktop, along with debugging guidance using the MCP Inspector.

Components

Details on collection schemas and resource templates for search queries and collection details.

Usage with Claude Desktop

Configuration instructions for integrating with Claude Desktop.

Logging

Logs server operations, requests, and errors to /tmp/typesense-mcp.log.

License

Licensed under the MIT License.