SingleStore

The SingleStore MCP Server lets you talk to SingleStore through the Model Context Protocol (MCP). With MCP, you can interact with SingleStore using natural language via compatible MCP clients, performing SQL operations, workspace actions, and management tasks in a streamlined way.

How to use

You connect an MCP client to the SingleStore MCP Server and start issuing natural language commands. Common tasks include running SQL queries, managing workspaces and notebooks, and inspecting user or organization details. If you ever want a quick tour, call the /help prompt from your chat to see available capabilities.

How to install

Prerequisites you need before installing the MCP server:

• Python 3.10 or newer

• uvx installed in your Python environment

• An MCP client such as Claude Desktop, Claude Code, Cursor, Windsurf, VS Code, Goose, Gemini CLI, LM Studio, or Qodo Gen

Standard MCP server configuration (stdio)

{
  "mcpServers": {
    "singlestore-mcp-server": {
      "command": "uvx",
      "args": [
        "singlestore-mcp-server",
        "start"
      ]
    }
  }
}

Docker configuration (stdio)

If you prefer running the MCP server inside Docker, you must provide an API key for OAuth when starting the container.

{
  "mcpServers": {
    "singlestore-mcp-server": {
      "command": "docker",
      "args": [
        "run", "-i", "--rm", "--init", "--pull=always",
        "-e", "MCP_API_KEY=your_api_key_here",
        "singlestore/mcp-server-singlestore"
      ]
    }
  }
}

Starting the server with Docker (optional note)

You can build the Docker image yourself if needed and then run it with the API key embedded in the environment. This approach is useful when you want to isolate the MCP server in a container.

Client-specific automatic setup

The following automatic setup commands initialize the SingleStore MCP Server for popular MCP clients. Use these from your terminal after you have the standard or Docker config in place.

Available tools

get_user_info

Retrieve details about the current user, including basic profile information.

organization_info

Retrieve details about the user's current organization.

choose_organization

Interactively select an organization from available options (when the API key is not set).

set_organization

Set the active organization by supplying an organization_id.

workspace_groups_info

Fetch details about the workspace groups accessible to the user.

workspaces_info

Fetch details about the workspaces within a specific workspace group.

resume_workspace

Resume a suspended workspace by providing the workspace_id.

list_starter_workspaces

List all starter workspaces available to the user.

create_starter_workspace

Create a new starter workspace using the provided configuration.

terminate_starter_workspace

Terminate a starter workspace by supplying the workspace_id.

list_regions

Return a list of all regions that support workspaces.

list_sharedtier_regions

Return a list of shared tier regions.

run_sql

Execute SQL against a connected workspace with given database and sql_query.

create_notebook_file

Create a new notebook in SingleStore Spaces with a name and optional content.

upload_notebook_file

Upload an existing notebook file to SingleStore Spaces.

create_job_from_notebook

Create a scheduled job based on a notebook and its configuration.

get_job

Retrieve details for a specific job by job_id.

delete_job

Delete a specific job by job_id.