GitHub - devlimelabs/meilisearch-ts-mcp (original) (raw)

Meilisearch MCP Server

A Model Context Protocol (MCP) server implementation for Meilisearch, enabling AI assistants to interact with Meilisearch through a standardized interface.

Features

• Index Management: Create, update, and delete indexes
• Document Management: Add, update, and delete documents
• Search Capabilities: Perform searches with various parameters and filters
• Settings Management: Configure index settings
• Task Management: Monitor and manage asynchronous tasks
• System Operations: Health checks, version information, and statistics
• Vector Search: Experimental vector search capabilities

Installation

Installing via Smithery

To install Meilisearch MCP Server for Claude Desktop automatically via Smithery:

npx -y @smithery/cli install @devlimelabs/meilisearch-ts-mcp --client claude

Manual Installation

1. Clone the repository:
   git clone https://github.com/devlimelabs/meilisearch-ts-mcp.git
   cd meilisearch-ts-mcp
2. Install dependencies:
3. Create a .env file based on the example:
4. Edit the .env file to configure your Meilisearch connection.

Docker Setup

The Meilisearch MCP Server can be run in a Docker container for easier deployment and isolation.

Using Docker Compose

The easiest way to get started with Docker is to use Docker Compose:

Start the Meilisearch MCP Server

docker-compose up -d

View logs

docker-compose logs -f

Stop the server

docker-compose down

Building and Running the Docker Image Manually

You can also build and run the Docker image manually:

Build the Docker image

docker build -t meilisearch-ts-mcp .

Run the container

docker run -p 3000:3000 --env-file .env meilisearch-ts-mcp

Development Setup

For developers who want to contribute to the Meilisearch MCP Server, we provide a convenient setup script:

Clone the repository

git clone https://github.com/devlimelabs-ts-mcp/meilisearch-ts-mcp.git cd meilisearch-ts-mcp

Run the development setup script

./scripts/setup-dev.sh

The setup script will:

1. Create a .env file from .env.example if it doesn't exist
2. Install dependencies
3. Build the project
4. Run tests to ensure everything is working correctly

After running the setup script, you can start the server in development mode:

Usage

Building the Project

Running the Server

Development Mode

Claude Desktop Integration

The Meilisearch MCP Server can be integrated with Claude for Desktop, allowing you to interact with your Meilisearch instance directly through Claude.

Automated Setup

We provide a setup script that automatically configures Claude for Desktop to work with the Meilisearch MCP Server:

First build the project

npm run build

Then run the setup script

node scripts/claude-desktop-setup.js

The script will:

1. Detect your operating system and locate the Claude for Desktop configuration file
2. Read your Meilisearch configuration from the .env file
3. Generate the necessary configuration for Claude for Desktop
4. Provide instructions for updating your Claude for Desktop configuration

Manual Setup

If you prefer to manually configure Claude for Desktop:

1. Locate your Claude for Desktop configuration file:
   • macOS: ~/Library/Application Support/Claude/claude_desktop_config.json
   • Windows: %APPDATA%\Claude\claude_desktop_config.json
   • Linux: ~/.config/Claude/claude_desktop_config.json
2. Add the following configuration (adjust paths as needed):

{ "mcpServers": { "meilisearch": { "command": "node", "args": ["/path/to/meilisearch-ts-mcp/dist/index.js"], "env": { "MEILISEARCH_HOST": "http://localhost:7700", "MEILISEARCH_API_KEY": "your-api-key" } } } }

1. Restart Claude for Desktop to apply the changes.
2. In Claude, type: "I want to use the Meilisearch MCP server" to activate the integration.

Cursor Integration

The Meilisearch MCP Server can also be integrated with Cursor, an AI-powered code editor.

Setting Up MCP in Cursor

1. Install and set up the Meilisearch MCP Server:
   git clone https://github.com/devlimelabs/meilisearch-ts-mcp.git
   cd meilisearch-ts-mcp
   npm install
   npm run build
2. Start the MCP server:
3. In Cursor, open the Command Palette (Cmd/Ctrl+Shift+P) and search for "MCP: Connect to MCP Server".
4. Select "Connect to a local MCP server" and enter the following details:
   • Name: Meilisearch
   • Command: node
   • Arguments: /absolute/path/to/meilisearch-ts-mcp/dist/index.js
   • Environment Variables:

   MEILISEARCH_HOST=http://localhost:7700  
   MEILISEARCH_API_KEY=your-api-key  

5. Click "Connect" to establish the connection.
6. You can now interact with your Meilisearch instance through Cursor by typing commands like "Search my Meilisearch index for documents about..."

Available Tools

The Meilisearch MCP Server provides the following tools:

Index Tools

• create-index: Create a new index
• get-index: Get information about an index
• list-indexes: List all indexes
• update-index: Update an index
• delete-index: Delete an index

Document Tools

• add-documents: Add documents to an index
• get-document: Get a document by ID
• get-documents: Get multiple documents
• update-documents: Update documents
• delete-document: Delete a document by ID
• delete-documents: Delete multiple documents
• delete-all-documents: Delete all documents in an index

Search Tools

• search: Search for documents
• multi-search: Perform multiple searches in a single request

Settings Tools

• get-settings: Get index settings
• update-settings: Update index settings
• reset-settings: Reset index settings to default
• Various specific settings tools (synonyms, stop words, ranking rules, etc.)

Task Tools

• list-tasks: List tasks with optional filtering
• get-task: Get information about a specific task
• cancel-tasks: Cancel tasks based on provided filters
• wait-for-task: Wait for a specific task to complete

System Tools

• health: Check the health status of the Meilisearch server
• version: Get version information
• info: Get system information
• stats: Get statistics about indexes

Vector Tools (Experimental)

• enable-vector-search: Enable vector search
• get-experimental-features: Get experimental features status
• update-embedders: Configure embedders
• get-embedders: Get embedders configuration
• reset-embedders: Reset embedders configuration
• vector-search: Perform vector search

License

MIT