ClickHouse MCP server

This is a small helper that connects your AI assistant to ClickHouse, a database used for storing and analyzing large amounts of data. Once it is set up, you can ask your AI questions in plain English and it will go and check the database for you. It is handy for anyone who works with data but does not want to write SQL by hand every time.

What this MCP server does

You ask your AI a question about your data, like which tables exist or how many rows are in a table. The AI passes that request to this helper program. The helper connects to your ClickHouse cluster and runs the query or lookup. The result comes back to the AI, which explains it to you in the chat. By default it only reads data, so it will not change anything unless someone turns writes on.

Click to zoom

What you can do with it

• List all databases on your ClickHouse cluster
• List the tables inside a database, with filters and paging
• Run read-only SQL queries and get the rows back
• Check what columns a query would return before running it
• Estimate how much data a query would read
• Run SQL against files and URLs using the built-in chDB engine

Try asking your AI

• “What databases do I have in ClickHouse?”
• “List the tables in the analytics database and show me their columns.”
• “Run this query and show me the first 20 rows: SELECT * FROM events LIMIT 20”
• “Before I run this, estimate how many rows it would read: SELECT count() FROM events WHERE id = 42”

What it gives back to you

You get answers in the chat, usually as a short explanation plus the rows or a table. Listing databases or tables comes back as a list you can read. Query results come back as rows with their column names. If something is wrong, you get the database's own error message so you can fix your question.

Before you start

Who it's for

Data analysts, support and operations people, and anyone who needs quick answers from a ClickHouse database without writing SQL themselves.