Keboola MCP Server

What does “Keboola” MCP Server do?

Keboola MCP Server acts as an open-source bridge between your Keboola project and modern AI tools. It connects AI assistants and MCP clients (such as Claude, Cursor, Windsurf, VS Code, and others) to the Keboola platform, exposing features like storage access, SQL transformations, component management, and job triggers as callable tools. This integration enables AI models and agents to query tables, manage configurations, execute jobs, and interact with metadata directly from their environment. By doing so, it streamlines development workflows, eliminates glue code, and ensures the right data and capabilities are available to AI agents when needed, enhancing productivity and enabling complex automation scenarios.

List of Prompts

List of Resources

List of Tools

Based on the repository’s features and available documentation, the following tools are provided by the Keboola MCP Server:

• Storage: Query tables directly and manage table or bucket descriptions within Keboola storage.
• Components: Create, list, and inspect extractors, writers, data apps, and transformation configurations.
• SQL: Create and execute SQL transformations using natural language.
• Jobs: Run components, trigger transformations, and retrieve job execution details.
• Metadata: Search, read, and update project documentation and object metadata.

Use Cases of this MCP Server

• Database Management: Directly query and manage tables or buckets in Keboola storage, allowing AI agents to fetch or modify project data.
• Codebase & Configuration Exploration: List, create, and inspect extractors, writers, and transformation configurations from AI tools, simplifying configuration management.
• Automated SQL Transformation: Use natural language to generate and execute SQL queries, enabling rapid transformation and analysis of stored data.
• Job Orchestration & Monitoring: Run components, orchestrate jobs, and retrieve execution histories, making it easy to automate and oversee ETL/data workflows.
• Metadata Handling: Search, read, and update project documentation and metadata to keep information organized and accessible for both humans and AI agents.

How to set it up

Windsurf

1. Ensure you have Python 3.10+ and uv installed.
2. Obtain your Keboola Storage API token and (if using a custom token) your workspace schema.
3. In Windsurf, locate the MCP configuration file.
4. Add the Keboola MCP Server entry using the following JSON snippet:

   {
     "mcpServers": {
       "keboola-mcp": {
         "command": "uv",
         "args": ["pip", "run", "--", "keboola-mcp-server"]
       }
     }
   }
   

5. Save the configuration file and restart Windsurf.
6. Verify server availability in the Windsurf MCP interface.

Securing API Keys (Windsurf)

{
  "mcpServers": {
    "keboola-mcp": {
      "command": "uv",
      "args": ["pip", "run", "--", "keboola-mcp-server"],
      "env": {
        "KBC_STORAGE_TOKEN": "${KBC_STORAGE_TOKEN}",
        "KBC_WORKSPACE_SCHEMA": "${KBC_WORKSPACE_SCHEMA}"
      },
      "inputs": {
        "KBC_STORAGE_TOKEN": "env",
        "KBC_WORKSPACE_SCHEMA": "env"
      }
    }
  }
}

Claude

1. Ensure Python 3.10+ and uv are installed.
2. Acquire the necessary Keboola credentials.
3. Open the Claude client MCP configuration.
4. Insert the Keboola MCP Server setup:

   {
     "mcpServers": {
       "keboola-mcp": {
         "command": "uv",
         "args": ["pip", "run", "--", "keboola-mcp-server"]
       }
     }
   }
   

5. Save and restart Claude.
6. Confirm the server is accessible from Claude.

Cursor

1. Install Python 3.10+ and uv.
2. Prepare your Keboola API token and workspace schema.
3. Open the Cursor MCP configuration file.
4. Add the following configuration:

   {
     "mcpServers": {
       "keboola-mcp": {
         "command": "uv",
         "args": ["pip", "run", "--", "keboola-mcp-server"]
       }
     }
   }
   

5. Save the file and restart Cursor.
6. Check for successful MCP server connection.

Cline

1. Make sure Python 3.10+ and uv are installed.
2. Gather required Keboola credentials.
3. Edit the MCP servers section in Cline’s configuration.
4. Add the Keboola MCP Server entry:

   {
     "mcpServers": {
       "keboola-mcp": {
         "command": "uv",
         "args": ["pip", "run", "--", "keboola-mcp-server"]
       }
     }
   }
   

5. Save configuration and restart Cline.
6. Verify proper server operation.

Note: Secure sensitive credentials like API tokens using environment variables as shown in the Windsurf example above.

How to use this MCP inside flows

Using MCP in FlowHunt

To integrate MCP servers into your FlowHunt workflow, start by adding the MCP component to your flow and connecting it to your AI agent:

Click on the MCP component to open the configuration panel. In the system MCP configuration section, insert your MCP server details using this JSON format:

{
  "keboola-mcp": {
    "transport": "streamable_http",
    "url": "https://yourmcpserver.example/pathtothemcp/url"
  }
}

Once configured, the AI agent is now able to use this MCP as a tool with access to all its functions and capabilities. Remember to change “keboola-mcp” to whatever the actual name of your MCP server is and replace the URL with your own MCP server URL.

Overview

My evaluation: The Keboola MCP Server provides a strong set of tools and clear setup instructions, but lacks documented prompt templates and explicit MCP resource definitions. Its focus on enabling AI agents to access complex data workflows is robust. Sampling and roots support are not documented. Overall, this is a highly practical and production-ready MCP, but with some documentation gaps for prompt/resources.

MCP Score