Meilisearch

Meilisearch is an MCP server that connects LLM clients directly to Meilisearch instances for automated full-text and semantic search operations. Designed for software developers, data engineers, and AI application builders, it allows assistants such as Claude or custom OpenAI agents to manage search indices, configure ranking parameters, and query structured records. Users can run full-text, vector-based, or hybrid searches across one or multiple indices using natural language commands. The server handles index creation with custom primary keys, paginated document retrieval, and bulk data ingestion. It also monitors instance health, tracks asynchronous indexing tasks, and inspects index metrics. In addition, the server provides tools to adjust searchable fields, update ranking rules, configure faceting, and manage API keys across local or remote instances using standard stdio communication with structured tool outputs.

Category: Databases & Data Stores

Tags: full-text search, indexing, meilisearch, search engine, semantic search

Visit Meilisearch

How to install and configure Meilisearch

1. Ensure you have Python 3.10 or higher installed and a running Meilisearch instance. 2. Install the package via pip or uvx: bash pip install meilisearch-mcp 3. Add the server configuration to your claude_desktop_config.json file: json { "mcpServers": { "meilisearch": { "command": "uvx", "args": ["-n", "meilisearch-mcp"] } } } 4. Optionally define connection environment variables such as MEILI_HTTP_ADDR and MEILI_MASTER_KEY. 5. Restart your client to begin querying Meilisearch through your assistant.

What you can do with Meilisearch

• Running multi-index search queries with complex filters, sorting, and faceted aggregations through conversational AI prompts. - Ingesting, updating, or paginating through large collections of JSON documents across development and production search indices. - Creating new search indices with specific primary keys or deleting obsolete collections directly from client interfaces. - Tuning search relevancy, typo tolerance, searchable attributes, and custom ranking rules without writing API scripts. - Monitoring background indexing jobs, tracking task status, and verifying Meilisearch server health metrics in real time.

Key facts

• Open Source
• https://github.com/meilisearch/meilisearch-mcp
• Databases & Data Stores, Web Search & Research
• full-text search, indexing, meilisearch, search engine, semantic search

Part of MCP Servers

What is Meilisearch MCP?

Meilisearch MCP is an official Model Context Protocol server that bridges AI assistants to Meilisearch search engines. It allows LLMs to interact with Meilisearch over stdio, enabling automated index management, document ingestion, settings configuration, and hybrid search queries directly through conversational workflows.

How do I install Meilisearch MCP?

You can install it using pip by running pip install meilisearch-mcp, or run it dynamically with uvx using uvx -n meilisearch-mcp. You then register the command in your MCP client configuration file, such as claude_desktop_config.json, pointing to the uvx command.

Which MCP clients work with Meilisearch MCP?

Meilisearch MCP works with any client that implements the Model Context Protocol over stdio. This includes Claude Desktop, custom OpenAI agents, containerized workflow engines like n8n, and developer testing utilities such as the official MCP Inspector.

Is Meilisearch MCP open source?

Yes, Meilisearch MCP is open source and distributed under the MIT License. The code repository is publicly hosted on GitHub, allowing developers to review, contribute, and adapt the Python server implementation for their own environments.