==================
/// MCP ///
/// MCP ///
==================
[server:online]
[protocol:ready]mcp-clickhouse
by ClickHouse
Connect ClickHouse to your AI assistants via an MCP (Model Context Protocol) server that exposes ClickHouse and chDB query tools plus a built-in health-check endpoint.
465
90
Open Source01
run_select_query
Execute a read-only SQL query against the configured ClickHouse cluster.
02
list_databases
Return a list of all databases available on the ClickHouse cluster.
03
list_tables
Return all tables within a specified ClickHouse database.
04
run_chdb_select_query
Execute a SQL query using chDB’s embedded OLAP engine, allowing direct querying of various data sources without ETL.
Installation
1. Clone the repository:
git clone https://github.com/ClickHouse/mcp-clickhouse.git
cd mcp-clickhouse2. (Recommended) Create & activate a Python virtual environment:
python -m venv .venv
source .venv/bin/activate # Windows: .venv\Scripts\activate3. Install the package and its dependencies:
pip install -U pip wheel
pip install -e . # Installs mcp-clickhouse in editable mode4. Configure connection to your ClickHouse cluster (environment variables):
export CLICKHOUSE_HOST=<host>
export CLICKHOUSE_PORT=<port>
export CLICKHOUSE_USER=<user>
export CLICKHOUSE_PASSWORD=<password>
# Optional: CLICKHOUSE_DATABASE=<db>5. Launch the MCP server (typical entry-point script):
mcp-clickhouse serve # or: python -m mcp_clickhouse.server6. Verify the server is running (default http://localhost:8000/health).
7. Integrate with your AI assistant by pointing it at the MCP endpoint you just started.
Documentation
License: Apache License 2.0
Updated 7/30/2025