Keboola

Keboola MCP Server is an open-source bridge that lets you connect Keboola data, SQL transformations, and workflow components to modern AI assistants and MCP clients. You can expose data, run transformations and jobs, and orchestrate data flows from your AI agents with no glue code needed.

How to use

Connect Keboola to your AI assistants and MCP clients to access data, run SQL transformations, trigger jobs, and manage workflows. Start by connecting to the remote MCP server or running a local instance, then configure your client to use the provided MCP endpoint. Once connected, you can ask your AI agent to query tables, run a SQL transformation, trigger a data extraction job, or manage workflows just like you would with any API-equipped tool.

How to install

Prerequisites: you need a Keboola project with admin rights, Python 3.10+ (or a suitable runtime for your chosen transport), and an MCP client (Claude, Cursor, Windsurf, VS Code, or another client). You may also use uvx to run the MCP server locally.

Choose your setup path and follow the steps below.

# Remote (recommended) setup via the remote MCP server
# No local installation required; configure your MCP client to connect to the remote endpoint
# Example remote endpoint for your region (you will replace YOUR_REGION):
# https://mcp.YOUR_REGION.keboola.com/mcp

# Local development or testing requires running the MCP server yourself (see options B-D below)

Additional setup options and steps

There are multiple ways to use Keboola MCP Server depending on your needs. You can opt for the remote hosted server for zero-setup usage, or run a local MCP server for full control and development. Each method provides a compatible MCP endpoint for your clients.

Available tools

Storage queries

Query storage tables directly and manage table or bucket descriptions.

Components management

Create, list and inspect extractors, writers, data apps, and transformation configurations.

SQL transformations

Create SQL transformations with natural language requests to transform data.

Jobs

Run components and transformations, and retrieve job execution details.

Flows and orchestration

Build Conditional Flows and Orchestrator Flows to manage workflows.

Data Apps

Create, deploy and manage Keboola Streamlit Data Apps displaying your data queries.

Metadata handling

Search, read, and update project documentation and object metadata using natural language.

Dev branches support

Work safely in development branches with scoped operations using branch IDs.