gateway

MCP.Pizza Chef: centralmind

CentralMind's tool sits in front of Postgres, MySQL, ClickHouse, Snowflake, SQL Server, BigQuery, Oracle, SQLite, or Elasticsearch and offers a reviewed, read-only set of questions instead of raw table access. It studies your tables once using a model you choose, and Google's free Gemini tier is enough for that step, then writes a settings file you can edit. Personal details can be stripped out automatically, and every request can be logged. It runs from Docker or a downloaded program and connects to Claude Desktop and Cursor.

Unmaintained · quiet since July 2025

Data

Use This MCP server To

Ask a question about last month's sales in plain words Let a teammate query the database without writing SQL Hide customer names and emails from anything the assistant sees Keep a record of every question asked against the database Connect a warehouse like Snowflake or BigQuery to my assistant

README

CentralMind Gateway: Create API or MCP Server in Minutes

🚀 Interactive Demo avialable here: https://centralmind.ai

What is Centralmind/Gateway

Simple way to expose your database to AI-Agent via MCP or OpenAPI 3.1 protocols.

docker run --platform linux/amd64 -p 9090:9090 \
  ghcr.io/centralmind/gateway:v0.2.18 start \
  --connection-string "postgres://db-user:db-password@db-host/db-name?sslmode=require"

This will run for you an API:

INFO Gateway server started successfully!         
INFO MCP SSE server for AI agents is running at: http://localhost:9090/sse 
INFO REST API with Swagger UI is available at: http://localhost:9090/

Which you can use inside your AI Agent:

Gateway will generate AI optimized API.

Why Centralmind/Gateway

AI agents and LLM-powered applications need fast, secure access to data. We're building an API layer that automatically generates secure, LLM-optimized APIs for your structured data.

gateway FAQ

Can it change or delete my data?

Not today. It is read-only, and adding or updating rows is listed as a future plan.

Do I need a key?

Only for the one-time setup step, where a model reads your tables and drafts the settings file. Google's free Gemini tier covers it.

Which databases are supported?

Postgres, MySQL, ClickHouse, Snowflake, SQL Server, BigQuery, Oracle, SQLite, and Elasticsearch.

Can I use this to let non-technical staff ask about company data?

That is the intent. You review the questions it generates first, and personal details can be redacted before anything reaches the assistant.

How hard is it to set up?

Expect a technical setup: Docker or a downloaded program, database connection details, and a settings file to review.

Which apps does it connect to?

Claude Desktop and Cursor are documented, and it is listed in several MCP directories including Smithery.

Is it actively developed?

It has been quiet since July 2025, so check the project's recent activity before relying on it for anything important.