meilisearch

The Meilisearch MCP Server lets you manage and interact with a Meilisearch instance through the Model Context Protocol. It exposes natural-language capable controls for connecting to one or more Meilisearch instances, managing indices and documents, performing searches, and adjusting settings—all via MCP clients like Claude or OpenAI agents. You can switch between different Meilisearch hosts on the fly, monitor health and task progress, and handle API keys and access control from conversations. The server supports index creation and management, document operations, multi-index searches, semantic/hybrid search (where available), and configurable ranking and filtering settings.

Once running, you can use your MCP client to: create indices with custom primary keys, add or update documents, run multi-index or semantic searches, configure ranking rules and facets, monitor indexing tasks, and manage API keys. It also offers health checks and system statistics to help you keep Meilisearch healthy and performing well. For deployment, you typically run the MCP server via uvx (as shown in Quick Start), or inside Docker, and connect it to your Meilisearch instance (via HTTP address and optional master key).

Prerequisites:

• Python 3.9+ installed on your system
• A running Meilisearch instance (local or remote)
• MCP-compatible client (Claude Desktop, OpenAI agents, etc.)

Option A: Install from PyPI (recommended)

pip install meilisearch-mcp

Option B: Install with uvx (recommended for MCP workflow)

uvx -n meilisearch-mcp

Option C: From Source (development)

# Clone repository
git clone https://github.com/meilisearch/meilisearch-mcp.git
cd meilisearch-mcp

# Create and activate virtual environment (example for Unix-like systems)
python -m venv .venv
source .venv/bin/activate

# Install in editable mode
uv pip install -e .

Using Docker

# Pull the latest image
docker pull getmeili/meilisearch-mcp:latest

# Run the container with required environment
docker run -it \
  -e MEILI_HTTP_ADDR=http://your-meilisearch:7700 \
  -e MEILI_MASTER_KEY=your-master-key \
  getmeili/meilisearch-mcp:latest

Optional: Docker Compose or Kubernetes workflows can wrap this for orchestration. If you plan to use n8n or Claude integrations, you can point them to the running MCP server endpoint via the configured STDIO or HTTP transport as supported by your MCP setup.

Environment variables you may configure:

• MEILI_HTTP_ADDR: URL of your Meilisearch instance (default http://localhost:7700)
• MEILI_MASTER_KEY: Optional master/key for authentication

Security note: The README indicates that editing hosts and API keys directly in chat is convenient for development but not recommended for production. Use proper secret management and follow MCP security best practices in production environments.

Common issues:

• Ensure Meilisearch is reachable at the configured MEILI_HTTP_ADDR before starting the MCP server.
• If using Docker, expose the correct ports and provide the appropriate MEILI_HTTP_ADDR and MEILI_MASTER_KEY.
• When switching between Meilisearch instances, update the connection settings to reflect the new host/address and credentials.

Troubleshooting tips:

• Check MCP server logs for connection errors to Meilisearch.
• Verify that the MCP client is sending valid MCP commands for index and document operations.
• Confirm that the MCP server has network access to the Meilisearch instance.