Using ClickHouse MCP server with LibreChat

This guide explains how to set up LibreChat with a ClickHouse MCP server using Docker.

This guide explains how to set up LibreChat with a ClickHouse MCP server using Docker and connect it to the ClickHouse example datasets.

1. Install docker

   You will need Docker to run LibreChat and the MCP server. To get Docker:

   1. Visit docker.com
   2. Download Docker desktop for your operating system
   3. Install Docker by following the instructions for your operating system
   4. Open Docker Desktop and ensure it is running

   
   

   For more information, see the Docker documentation.

2. Clone the LibreChat repository

   Open a terminal (command prompt, terminal or PowerShell) and clone the LibreChat repository using the following command:

   git clone https://github.com/danny-avila/LibreChat.git
   cd LibreChat
   
3. Create and edit the .env file

   Copy the example configuration file from .env.example to .env:

   cp .env.example .env
   

   Open the .env file in your favorite text editor. You will see sections for many popular LLM providers, including OpenAI, Anthropic, AWS bedrock etc, for example:

   .venv

   #============#
   # Anthropic  #
   #============#
   ANTHROPIC_API_KEY=user_provided
   # ANTHROPIC_MODELS=claude-opus-4-20250514,claude-sonnet-4-20250514,claude-3-7-sonnet-20250219,claude-3-5-sonnet-20241022,claude-3-5-haiku-20241022,claude-3-opus-20240229,claude-3-sonnet-20240229,claude-3-haiku-20240307
   # ANTHROPIC_REVERSE_PROXY=
   

   Replace user_provided with your API key for the LLM provider you want to use.

   Using a local LLM

   If you don’t have an API key you can use a local LLM like Ollama. You’ll see how to do this later in step “Install Ollama”. For now don’t modify the .env file and continue with the next steps.

4. Create a librechat.yaml file

   Run the following command to create a new librechat.yaml file:

   cp librechat.example.yaml librechat.yaml
   

   This creates the main configuration file for LibreChat.

5. Add ClickHouse MCP server to Docker compose

   Next we’ll add the ClickHouse MCP server to the LibreChat Docker compose file so that the LLM can interact with the ClickHouse SQL playground.

   Create a file called docker-compose.override.yml and add the following configuration to it:

   docker-compose.override.yml

   services:
     api:
       volumes:
         - ./librechat.yaml:/app/librechat.yaml
     mcp-clickhouse:
       image: mcp/clickhouse
       container_name: mcp-clickhouse
       ports:
         - 8001:8000
       extra_hosts:
         - "host.docker.internal:host-gateway"
       environment:
         - CLICKHOUSE_HOST=sql-clickhouse.clickhouse.com
         - CLICKHOUSE_USER=demo
         - CLICKHOUSE_PASSWORD=
         - CLICKHOUSE_MCP_SERVER_TRANSPORT=sse
         - CLICKHOUSE_MCP_BIND_HOST=0.0.0.0
   

   If you want to explore your own data, you can do so by using the host, username and password of your own ClickHouse Cloud service.

   Get started with ClickHouse Cloud

   If you don’t have a Cloud account yet, get started with ClickHouse Cloud today and receive $300 in credits. At the end of your 30-day free trial, continue with a pay-as-you-go plan, or contact us to learn more about our volume-based discounts. Visit our pricing page for details.

6. Configure MCP server in librechat.yaml

   Open librechat.yaml and place the following configuration at the end of the file:

   mcpServers:
     clickhouse-playground:
       type: sse
       url: http://host.docker.internal:8001/sse
   

   This configures LibreChat to connect to the MCP server running on Docker.

   Find the following line:

   librechat.yaml

   socialLogins: ['github', 'google', 'discord', 'openid', 'facebook', 'apple', 'saml']
   

   For simplicity, we will remove the need to authenticate for now:

   librechat.yaml

   socialLogins: []
   
7. Add a local LLM using Ollama (optional)

   Install Ollama

   Go to the Ollama website and install Ollama for your system.

   Once installed, you can run a model like this:

   ollama run qwen3:32b
   

   This will pull the model to your local machine if it isn’t present.

   For a list of models see the Ollama library

   Configure Ollama in librechat.yaml

   Once the model has downloaded, configure it in librechat.yaml:

   librechat.yaml

   custom:
     - name: "Ollama"
       apiKey: "ollama"
       baseURL: "http://host.docker.internal:11434/v1/"
       models:
         default:
           [
             "qwen3:32b"
           ]
         fetch: false
       titleConvo: true
       titleModel: "current_model"
       summarize: false
       summaryModel: "current_model"
       forcePrompt: false
       modelDisplayLabel: "Ollama"
   
8. Start all services

   From the root of the LibreChat project folder, run the following command to start the services:

   docker compose up
   

   Wait until all services are fully running.

9. Open LibreChat in your browser

   Once all services are up and running, open your browser and go to http://localhost:3080/

   Create a free LibreChat account if you don’t yet have one, and sign in. You should now see the LibreChat interface connected to the ClickHouse MCP server, and optionally, your local LLM.

   From the chat interface, select clickhouse-playground as your MCP server:

   You can now prompt the LLM to explore the ClickHouse example datasets. Give it a go:

   Prompt

   What datasets do you have access to?
   

If the MCP server option doesn’t appear in the LibreChat UI, check that the proper permissions are set in your librechat.yaml file.

If use is set to false in the mcpServers section under interface, the MCP selection dropdown won’t appear in chat:

librechat.yaml

interface:
  mcpServers:
    use: true
    share: false
    create: false
    public: false