MCP server for Meilisearch, enabling LLMs to search and manage database indexes.
1.2K
A high-performance Go implementation of the Model Context Protocol (MCP) for Meilisearch. This server enables AI assistants and desktop environments (like Claude Desktop, Cursor, Zed, etc.) to connect directly with your Meilisearch database to query, manage, and inspect search schemas, documents, and settings.
meilisearch-go SDK (v0.36.3) including index administration, settings, keys, tasks, custom webhooks, experimental controls, chat workspace integration, and network topology..md), allowing MCP-compatible LLMs to fetch and inspect the precise API specification on demand.meilisearch:// URI scheme.docker run --rm -p 8080:8080 \
-e MEILI_HOST=http://host.docker.internal:7700 \
-e MEILI_API_KEY=masterKey \
ja7adr/meilisearch-mcp:latest
MEILI_HOST (default: http://localhost:7700)MEILI_API_KEY (optional)MCP_ADDR (default: :8080)MCP_POOL_SIZE (default: 100)MCP_POOL_DURATION (default: 5m)MCP_RATE_LIMIT_RPS (default: 300)MCP_DEBUG (1 to enable debug logging)services:
mcp:
image: ja7adr/meilisearch-mcp:latest
build: .
environment:
MEILI_HOST: http://meili:7700
MEILI_API_KEY: ${MEILI_API_KEY:-masterKey}
MCP_RATE_LIMIT_RPS: 200
ports:
- "8080:8080"
depends_on:
- meili
meili:
image: getmeili/meilisearch:latest
environment:
MEILI_MASTER_KEY: masterKey
ports:
- "7700:7700"
Content type
Image
Digest
sha256:c3b666d5a…
Size
17.5 MB
Last updated
3 months ago
docker pull ja7adr/meilisearch-mcp