Typesense Search MCP Server

SearchTypeScript

Access and search Typesense collections from AI models

Available Tools

typesense_query

Search for documents in Typesense collections with powerful filtering options. Accepts query text, collection name, search fields, filters, sort options, and limit parameters. Returns matching documents with relevance scores.

typesense_get_document

Retrieve specific documents by ID from collections. Requires collection name and document ID as input. Returns the complete document data.

typesense_collection_stats

Get statistics about a Typesense collection. Requires collection name as input. Returns collection metadata, document count, and schema information.

Typesense Search provides AI models with powerful search capabilities through the Typesense search engine. This MCP server enables LLMs to discover, search, and analyze data stored in Typesense collections with advanced filtering, sorting, and retrieval options. With Typesense Search, AI assistants can perform semantic searches across your data collections, retrieve specific documents by ID, and analyze collection statistics. The server exposes Typesense collections as resources that can be browsed and queried through a simple interface.

Getting Started with Typesense Search

Typesense Search provides a Model Context Protocol (MCP) server that connects AI models to your Typesense search engine. This allows AI assistants to search, retrieve, and analyze data from your Typesense collections.

Prerequisites

Node.js installed on your system
A running Typesense server instance
Typesense API key with appropriate permissions

Installation Options

Via npm

You can install the Typesense MCP server globally:

npm install -g typesense-mcp-server

Or as a local project dependency:

npm install typesense-mcp-server

Via mcp-get

If you use the MCP package manager:

npx @michaellatman/mcp-get@latest install typesense-mcp-server

Configuration

The Typesense MCP server requires configuration to connect to your Typesense instance. You'll need to provide:

Typesense API key
Host information (URL/IP address)
Port number
Protocol (http/https)

These can be configured through environment variables or directly in your client configuration.

Using with Claude Desktop

To use Typesense Search with Claude Desktop, add the server configuration to your Claude Desktop config file:

On MacOS: ~/Library/Application Support/Claude/claude_desktop_config.json
On Windows: %APPDATA%/Claude/claude_desktop_config.json

Add the following configuration, adjusting the environment variables to match your Typesense setup:

{
  "mcpServers": {
    "typesense-mcp-server": {
      "command": "typesense-mcp-server",
      "env": {
        "TYPESENSE_API_KEY": "your-api-key",
        "TYPESENSE_HOST": "localhost",
        "TYPESENSE_PORT": "8108",
        "TYPESENSE_PROTOCOL": "http"
      }
    }
  }
}

Docker Support

Typesense Search can also be run as a Docker container:

docker run -p 3000:3000 \
  -e TYPESENSE_API_KEY=your-api-key \
  -e TYPESENSE_HOST=your-typesense-host \
  -e TYPESENSE_PORT=8108 \
  -e TYPESENSE_PROTOCOL=http \
  typesense-mcp-server

Usage

Once configured, AI models can access your Typesense collections through:

Resources: Browse collections via typesense:// URIs
Tools: Execute searches, retrieve documents, and get collection statistics
Prompts: Analyze collections and get search suggestions

The AI will automatically discover available collections and can search through them based on your queries. You can ask the AI to:

Search for specific information across collections
Retrieve detailed information about particular documents
Analyze collection structure and contents
Get statistics about your data

Troubleshooting

If you encounter issues:

Verify your Typesense server is running and accessible
Check that your API key has appropriate permissions
Ensure the host, port, and protocol settings are correct
Look for error messages in the server logs

For development or debugging, you can run the server with additional logging:

DEBUG=typesense-mcp-server:* typesense-mcp-server