MCP For Claude Marketplace
ClickHouse

ClickHouse MCP Connector for Claude

A+

Query and manage ClickHouse databases — execute analytical SQL, monitor replication status, and manage schemas directly from your AI agent.

4 tools Official Updated Oct 1, 2026 Official Vinkius Partner
More Details Connect to Claude

Connect your ClickHouse instance to any AI agent to perform lightning-fast OLAP queries and manage your data infrastructure through natural language.

What you can do

  • Analytical Queries — Execute read-only SELECT queries with automatic safety limits using the select_query tool.
  • Schema Management — Create, alter, or drop tables and databases using the execute_query action for mutating operations.
  • Health Monitoring — Instantly verify server availability with ping and check replication lag across clusters with replicas_status.
  • Performance Tuning — Pass specific settings like max_rows_to_read or max_execution_time to optimize resource usage.

How it works

  1. Subscribe to this server
  2. Enter your ClickHouse URL, Username, and Password
  3. Start querying your big data from Claude, Cursor, or any MCP-compatible client

Who is this for?

  • Data Engineers — quickly inspect table schemas and row counts without switching to a SQL IDE.
  • Analysts — run complex aggregations and get instant summaries of large datasets.
  • DevOps — monitor cluster health and replication status during maintenance or incidents.
olapsqlbig-datadatabase-managementanalytics

4 tools expose this connector's capabilities to your AI agent.

ping

Check ClickHouse server health

select_query

Automatically enforces readonly=1. Execute a read-only SELECT query on ClickHouse

execute_query

Execute a mutating query on ClickHouse

replicas_status

Check ClickHouse replica delay

See how to talk to your AI agent using ClickHouse.

Check if the ClickHouse server is online and reachable.

I've called the `ping` endpoint and the server is healthy and responding correctly.

Run a select_query to find the top 5 largest tables in the 'system' database.

I've executed the query on the system.parts table. The largest tables are: 'hits' (50GB), 'events' (42GB), 'logs' (30GB)... Would you like to see the full breakdown?

Use execute_query to create a new table called 'mcp_test' with an id column.

The `execute_query` action was successful. The table 'mcp_test' has been created in the default database with the specified schema.

Yes. The `select_query` tool is strictly read-only and automatically enforces the `readonly=1` setting, ensuring that only SELECT, SHOW, and DESCRIBE operations are permitted.

Related Connectors

StarRocks

StarRocks MCP

10 tools Official

High-performance analytical database — manage clusters, tables, and query data via AI.

A+ View details →
RisingWave (Streaming Database)

RisingWave (Streaming Database) MCP

6 tools Official

Manage your RisingWave streaming database — execute SQL, ingest events, and monitor tables, materialized views, sources, and sinks directly from any AI agent.

A+ View details →
Snowflake

Snowflake MCP

11 tools Official

Execute SQL queries, manage databases, and analyze data on Snowflake with AI agents.

A+ View details →
Sumo Logic

Sumo Logic MCP

9 tools Official

Connect your AI to Sumo Logic. Orchestrate log searches, monitor active collectors, and analyze system events efficiently.

A+ View details →