AI Features

Keboola Model Context Protocol (MCP) Server

Connect your MCP clients and AI assistants to your Keboola Project and give them the powers of a Keboola Expert user:

Cursor - Direct deeplink installation
Claude - Organization-level integration
ChatGPT - Custom connector for Plus/Pro users
Windsurf - Manual configuration
VS Code - Agent mode with MCP servers
Make - Agent mode with MCP servers
Other clients - Remote server connection

Query data, create transformations, write SQL queries, and build your pipelines — all with no clicks required.

Keboola’s MCP Server brings powerful AI agents like Claude and Cursor directly into your data workflows. Just describe what you need in plain language to:

Search and explore your data.
Set up and manage workflows — no code required.
Build SQL queries and transformations using AI.
Launch and monitor jobs in real time.
Automatically document everything — down to the column level.

Business Use Cases

Analyze customer orders, segment users, or monitor campaign performance — just by describing what you need, your assistant will query the data in your project directly.
Instantly get data from any system (CRM, ERP, ads) without touching the UI.
Update and run data pipelines before key meetings — no technical help required.
Automatically generate clear, human-readable descriptions down to the column level for full data transparency.

Claude debug error flow in MCP Server

Connecting to Keboola’s MCP Server

Keboola MCP Server is hosted on every multi-tenant stack and supports OAuth authentication. You can use the remote server in any AI Assistant that supports remote Streamable HTTP connection and OAuth authentication. Streamable HTTP is the recommended transport method, providing bidirectional streaming for improved performance and reliability.

Remote Server Setup

In case your AI assistant supports remote connection, you can connect to Keboola’s MCP Server by following these steps:

Obtain the remote server URL of the stack https://mcp.<YOUR_REGION>.keboola.com/mcp.
- Available stack URLs:
  - https://mcp.keboola.com/mcp
  - https://mcp.us-east4.gcp.keboola.com/mcp
  - https://mcp.eu-central-1.keboola.com/mcp
  - https://mcp.north-europe.azure.keboola.com/mcp
  - https://mcp.europe-west3.gcp.keboola.com/mcp
- You can find the url in your Keboola Project Settings under the tab MCP Server
  - In there you can also find specific instructions for various clients.
Copy the server URL and paste it into your AI assistant’s settings.
Once you save the settings and refresh your AI assistant, you will be prompted to authenticate with your Keboola account and select the project you want to connect to.

Using with Claude Desktop

In Claude, go to Settings → Connectors
Click “Add custom connector”
Give the connector a name (Keboola) and paste in your Integration URL
- https://mcp.<YOUR_REGION>.keboola.com/mcp
Click “Add”
You’ll be prompted to authenticate with your Keboola account and select the project you want to connect to.

Via mcp-remote adapter (fallback)

For MCP clients that don’t yet support native remote (OAuth) connections, you can bridge to the remote server with the mcp-remote adapter.

NOTE: This method requires you to have Node.js installed on your computer.

Open the Claude menu on your computer and select “Settings…”
Click on “Developer” in the left-hand bar of the Settings pane, and then click on “Edit Config”
Paste the following JSON into the config file and save.

{
  "mcpServers": {
    "keboola": {
      "command": "npx",
      "args": [
        "mcp-remote",
        "https://mcp.<YOUR_REGION>.keboola.com/mcp"
      ]
    }
  }
}

Restart Claude Desktop, you’ll be prompted to authenticate with your Keboola account and select the project you want to connect to.

Using with ChatGPT

ChatGPT Plus, Pro, Business, and Enterprise users can connect to Keboola’s MCP Server using custom connectors. Follow these steps to set up the integration:

Step 1: Access ChatGPT Settings

Log in to your ChatGPT Plus or Pro account
Click on your profile icon in the bottom-left corner of the screen
From the menu that appears, select Settings

Step 2: Navigate to Connectors

In the settings window, select the Connectors tab from the left-hand sidebar
Open Advanced settings and toggle on Developer mode
Read and acknowledge the warning that appears - this mode allows you to create custom connectors
Back on the Connectors page, click Create

Step 3: Configure the New Connector

A new window will open for the “New Connector” settings
Name: Give your connector a name, such as “Keboola”
Description: Provide a brief description (optional)
MCP Server URL: This is the most important part. You need the specific URL for your Keboola MCP Server:
- Open a new tab and go to your Keboola project settings
- Navigate to Users & Settings > MCP Server
- Copy the MCP Server URL provided there (e.g., https://mcp.us-east4.gcp.keboola.com/mcp)
- Paste this URL into the MCP Server URL field in ChatGPT
Authentication: The authentication method will be set to OAuth automatically
Check the box to confirm you understand the “beta” nature of custom connectors
Click Create

Step 4: Authorize the Connection

After clicking “Create,” you will be redirected to the Keboola platform to authorize the connection
Sign in to your Keboola account if prompted
Select the specific Keboola project you want to connect to
Once authorized, you will be redirected back to ChatGPT
A confirmation message “Keboola is now connected” will appear at the top of the screen

Step 5: Use the Connector in a Chat

In a chat, enable the Keboola connector for the conversation via the message composer’s tools/connectors menu (the exact control varies by ChatGPT version). Once it’s enabled, you can ask questions about your Keboola data and ChatGPT will use the connector — for example: “What data tables are in my project?” or “Show me the latest job runs.”

Using with Cursor

Click the button related to your region below:

Stack (Region)	Cursor Deeplink
US Virginia AWS (default)
US Virginia GCP (us-east4)
EU Frankfurt AWS (eu-central-1)
EU Ireland Azure (north-europe)
EU Frankfurt GCP (europe-west3)

Alternatively, you can:

Navigate to Keboola project settings, click Users & Settings > MCP Server.
Click the Cursor tab.
Click the “Install In Cursor” button.
You’ll be prompted to login into your Keboola account and select the project you want to connect to.

Using with Windsurf

Windsurf supports MCP through its native integration with Cascade. Add Keboola’s MCP Server with a manual configuration:

Open the Windsurf settings and navigate to Cascade > Plugins
Click on “Add MCP Plugin” or edit the mcp_config.json file directly
Add the Keboola MCP Server configuration:

{
  "mcpServers": {
    "keboola": {
      "serverUrl": "https://mcp.<YOUR_REGION>.keboola.com/mcp"
    }
  }
}

Replace <YOUR_REGION> with your specific stack URL from the available stack URLs
Press the refresh button after adding the configuration
You’ll be prompted to authenticate with your Keboola account and select the project

For detailed instructions and troubleshooting, see the Windsurf MCP documentation.

Using with VS Code

VS Code supports MCP servers through GitHub Copilot’s agent mode. Follow these steps to set up Keboola’s MCP Server:

Prerequisites

VS Code with GitHub Copilot extension installed
GitHub Copilot subscription (Pro, Business, or Enterprise)

Setup Instructions

Open VS Code and ensure you have the latest version of the GitHub Copilot extension
Open the Command Palette (Ctrl+Shift+P / Cmd+Shift+P) and run “MCP: Add Server”
This will create or open your MCP configuration file (mcp.json)
Add the Keboola MCP Server configuration (VS Code connects to remote servers natively over Streamable HTTP — no adapter needed):

{
  "servers": {
    "keboola": {
      "type": "http",
      "url": "https://mcp.<YOUR_REGION>.keboola.com/mcp"
    }
  }
}

Replace <YOUR_REGION> with your specific stack URL from the available stack URLs
Save the configuration file
Restart VS Code or run “MCP: Restart Servers” from the Command Palette
You’ll be prompted to authenticate with your Keboola account and select the project

Using MCP Tools in VS Code

Open the Chat view and enable agent mode
Select the Tools button to see available Keboola tools
Use # in your chat to reference specific tools or resources
The available Keboola tools will appear in the tools list for the agent

For detailed setup and troubleshooting, see the VS Code MCP documentation.

Using with MAKE

Create your scenario in MAKE.
Add the MCP Client module to the scenario.
Navigate to Keboola project settings, click Users & Settings > MCP Server, and open the MAKE tab.
Copy the integration URL displayed there and enter it as the server URL in the MCP Client module. MAKE loads the available Keboola tools automatically.

Available Tools

Keboola’s MCP Server comes with a rich set of tools your AI client (like Claude or Cursor) can use to interact with your data environment — just by asking in plain language. Don’t worry about remembering command names — your AI client handles that. Just describe what you want to do.

Components & Transformations – Create, edit, and launch them with natural language.
Storage – Browse, edit, and document buckets, tables, and columns.
SQL – Run and manage SQL queries.
Semantic layer – Explore the project’s semantic models and validate queries against them.
Jobs – Start, monitor, and debug execution flows.
Flows – Create and manage flows (including conditional flows) that orchestrate your components.
Data Apps – Create, deploy, and manage Streamlit and Python/JS data apps.
Search & Discovery – Find components, configurations, and objects across your project.
Project & OAuth – Read project info and set up OAuth authorizations for components.
Documentation – Search official Keboola docs from within your AI chat.

Restricting Tool Access

When using the remote MCP server, you may want to limit which tools are available to AI agents. This is useful for:

AI Agent Restrictions: Limiting what actions an AI agent (like Devin or Cursor) can perform in your project
Compliance and Security: Enforcing data governance policies by restricting write operations
Customer-Specific Access: Creating tailored access profiles for different use cases

When connecting via the Streamable HTTP transport, you control which tools are available to clients using HTTP headers.

Authorization Headers

The following HTTP headers control tool access:

Header	Description	Example Value
`X-Allowed-Tools`	Comma-separated list of tool names to allow. Only these tools will be available.	`get_configs,get_buckets,query_data`
`X-Disallowed-Tools`	Comma-separated list of tool names to exclude. These tools will be removed from the available set.	`create_config,run_job`
`X-Read-Only-Mode`	When set to `true`, `1`, or `yes`, restricts access to read-only tools only.	`true`

These headers are set by the client (e.g., your AI agent integration or custom MCP client) when making HTTP requests to the MCP server. Refer to your MCP client’s documentation for how to configure custom HTTP headers.

Filter Behavior

When multiple headers are present, filters are applied in the following order:

Allowed tools filter: If X-Allowed-Tools is specified, only those tools are initially available.
Read-only intersection: If X-Read-Only-Mode is enabled, the available tools are intersected with the read-only tools set.
Disallowed exclusion: Tools listed in X-Disallowed-Tools are removed from the final set.

Empty headers are treated as no restriction/exclusion (backward compatible behavior).

Read-Only Tools

The following tools are classified as read-only (they do not modify data). The live set may grow over time — see TOOLS.md in the server repo for the current annotations:

Category	Tools
Components	`get_configs`, `get_components`, `get_config_examples`, `run_sync_action`
Flows	`get_flows`, `get_flow_examples`, `get_flow_schema`
Storage	`get_buckets`, `get_tables`
SQL	`query_data`
Semantic	`get_semantic_context`, `get_semantic_schema`, `search_semantic_context`, `validate_semantic_query`
Data Apps	`get_data_apps`
Jobs	`get_jobs`
Search	`search`, `find_component_id`
Project	`get_project_info`
Documentation	`docs_query`

Examples

AI Agent Restrictions: When integrating AI agents (like Devin, Cursor, or custom agents) with your Keboola project, you may want to limit their capabilities. For example, allowing an agent to query and explore data but preventing it from creating or modifying configurations:

X-Read-Only-Mode: true

Compliance and Security: For environments with strict data governance requirements, you can create customer-specific access profiles. For example, allowing only specific tools while explicitly blocking others:

X-Allowed-Tools: get_buckets,get_tables,query_data,search
X-Disallowed-Tools: run_job

Combined Restrictions: You can combine all three headers for fine-grained control. For example, to allow only a subset of read-only tools:

X-Allowed-Tools: get_configs,get_buckets,get_tables,query_data,create_config
X-Read-Only-Mode: true
X-Disallowed-Tools: query_data

This configuration would result in only get_configs, get_buckets, and get_tables being available (the intersection of allowed and read-only, minus the disallowed).

Development Branches

To keep an AI agent’s changes off production, you can scope the MCP server to a development branch:

Remote server: send the X-Branch-Id HTTP header with your branch ID (set by the client, like the tool-authorization headers above).
Local server: set the KBC_BRANCH_ID environment variable.

Without it, the server operates on the production branch.

Running the MCP Server Locally

While MCP clients like Cursor or Claude typically manage the MCP server automatically, you might want to run the Keboola MCP Server locally for development, testing, or when using a custom client. You can run it via Docker or via the uv/uvx command.

Using Docker (recommended)

For a consistent and isolated environment, running the Keboola MCP Server via Docker is often the recommended approach for local execution, especially if you don’t want to manage Python environments directly or are integrating with clients that can manage Docker containers.

Before proceeding, ensure you have Docker installed on your system. You can find installation guides on the official Docker website.

Pull the latest image:
Terminal window
```
docker pull keboola/mcp-server:latest
```

Run the Docker container:

For Snowflake users:

docker run -it --rm \
  -e KBC_STORAGE_TOKEN="YOUR_KEBOOLA_STORAGE_TOKEN" \
  -e KBC_WORKSPACE_SCHEMA="YOUR_WORKSPACE_SCHEMA" \
  keboola/mcp-server:latest \
  --api-url https://connection.YOUR_REGION.keboola.com

Replace YOUR_KEBOOLA_STORAGE_TOKEN, YOUR_WORKSPACE_SCHEMA, and https://connection.YOUR_REGION.keboola.com with your actual values.

For BigQuery users (requires volume mount for credentials):

# Ensure your Google Cloud credentials JSON file is accessible
docker run -it --rm \
  -e KBC_STORAGE_TOKEN="YOUR_KEBOOLA_STORAGE_TOKEN" \
  -e KBC_WORKSPACE_SCHEMA="YOUR_WORKSPACE_SCHEMA" \
  -e GOOGLE_APPLICATION_CREDENTIALS="/creds/credentials.json" \
  -v /local/path/to/your/credentials.json:/creds/credentials.json \
  keboola/mcp-server:latest \
  --api-url https://connection.YOUR_REGION.keboola.com

Replace placeholders and ensure /local/path/to/your/credentials.json points to your actual credentials file on your host machine.

The --rm flag ensures the container is removed when it stops. The server inside Docker will typically listen on stdio by default, which is suitable for clients that can invoke and manage Docker commands.

Example: Configuring Cursor IDE to use Docker for Keboola MCP Server:

If your MCP client (like Cursor) supports defining a Docker command for an MCP server, the configuration might look like this:

{
  "mcpServers": {
    "keboola": {
      "command": "docker",
      "args": [
        "run",
        "-it",
        "--rm",
        "-e", "KBC_STORAGE_TOKEN",
        "-e", "KBC_WORKSPACE_SCHEMA",
        "keboola/mcp-server:latest",
        "--api-url", "https://connection.YOUR_REGION.keboola.com"
      ],
      "env": {
        "KBC_STORAGE_TOKEN": "YOUR_KEBOOLA_STORAGE_TOKEN",
        "KBC_WORKSPACE_SCHEMA": "YOUR_WORKSPACE_SCHEMA"
      }
    }
  }
}

Note

Ensure Docker is running on your system.
Replace placeholders like YOUR_KEBOOLA_STORAGE_TOKEN, YOUR_WORKSPACE_SCHEMA, and the Keboola API URL.
The client (Cursor) passes the KBC_STORAGE_TOKEN and KBC_WORKSPACE_SCHEMA from its env block to the docker run command through the -e flags. The --api-url is passed directly as an argument to the keboola/mcp-server entrypoint.

Using the uv command

The primary way to run the server locally without Docker is by using uv or uvx to execute the keboola_mcp_server package. More information about the server is available in its Keboola MCP Server GitHub repository. Make sure you have Python 3.10+ and uv installed.

Set up environment variables:
Before running the server, you need to configure the following environment variables:
- KBC_STORAGE_TOKEN: Your Keboola Storage API token.
- KBC_WORKSPACE_SCHEMA: Your Keboola project’s workspace schema (for SQL queries).
- KBC_STORAGE_API_URL: Your Keboola instance API URL (e.g., https://connection.keboola.com or https://connection.YOUR_REGION.keboola.com).
- KBC_BRANCH_ID (optional): a development branch ID to scope operations to; defaults to the production branch.
Refer to the Keboola Tokens and Keboola workspace manipulation for detailed instructions on obtaining these values.

1.1. Additional Setup for BigQuery Users
If your Keboola project uses BigQuery as its backend, you will also need to set up the GOOGLE_APPLICATION_CREDENTIALS environment variable. This variable should point to the JSON file containing your Google Cloud service account key that has the necessary permissions to access your BigQuery data.

Example:
GOOGLE_APPLICATION_CREDENTIALS="/path/to/your/credentials.json"
Run the server:

uvx keboola_mcp_server --api-url $KBC_STORAGE_API_URL

The KBC_STORAGE_API_URL was set as an environment variable but can also be provided manually via the --api-url flag. The command starts the server communicating via stdio. To run the server in Streamable HTTP mode (listening on a network host/port such as localhost:8000), pass the appropriate flags to keboola_mcp_server. For day-to-day use with clients like Claude or Cursor you usually do not need to run this command manually, as they handle the server lifecycle.

Connecting a client to a localhost instance

When you run the Keboola MCP Server manually, it will typically listen on stdio or on a specific HTTP port if configured for Streamable HTTP.

stdio-based clients: Configure the client application to launch the local keboola_mcp_server executable and communicate over standard input/output.
Streamable HTTP-based clients: If you start the server in HTTP mode, your client should connect to the specified host and port (e.g., http://localhost:8000/mcp?storage_token=XXX&workspace_schema=YYY).

Example: connecting Cursor IDE to a local uvx instance

If you are running the Keboola MCP Server locally using uvx, you can configure Cursor IDE to connect to this local instance. This is useful for development or testing with a custom server build.

Open Cursor settings
Navigate to the MCP section within settings
Add or configure your Keboola project Provide your KBC_STORAGE_TOKEN, KBC_WORKSPACE_SCHEMA and the API URL.

Example mcp_servers.json snippet:

{
  "mcpServers": {
    "keboola": {
      "command": "uvx",
      "args": [
        "keboola_mcp_server",
        "--api-url", "https://connection.YOUR_REGION.keboola.com"
      ],
      "env": {
        "KBC_STORAGE_TOKEN": "your_keboola_storage_token",
        "KBC_WORKSPACE_SCHEMA": "your_workspace_schema"
      }
    }
  }
}

You can use this link to get the above configuration template into your Cursor:

Programmatic Integration

Beyond ready-made clients, you can integrate the Keboola MCP Server directly into your own code and AI agent frameworks. This unlocks fully automated data workflows driven by natural-language instructions.

Claude Messages API with MCP Connector (Beta)

Anthropic offers a beta feature, the MCP connector, which enables you to connect to remote MCP servers (such as the Keboola MCP Server) directly through Claude’s Messages API. This method bypasses the need for a separate, standalone MCP client if you are already using the Claude Messages API.

Key features of this integration

Direct API Calls: You configure connections to MCP servers by including the mcp_servers parameter in your API requests to Claude, and enable tools via an mcp_toolset in the tools array.
Beta header: The connector is a beta feature — send anthropic-beta: mcp-client-2025-11-20 (the earlier mcp-client-2025-04-04 version is deprecated).
Tool Calling: The primary MCP functionality currently supported through this connector is tool usage.
Accessibility: The target MCP server needs to be publicly accessible over HTTP.

This approach can simplify your architecture if you’re building applications that programmatically interact with Claude and need to leverage MCP-enabled tools without managing an additional client layer.

For complete details, API examples, and configuration options, please consult the official Anthropic MCP connector documentation.

OpenAI Agents SDK (Python)

The OpenAI Agents SDK ships with first-class MCP support. Simply start the Keboola MCP Server (locally via uvx or remotely over Streamable HTTP) and register it with the SDK:

from agents import Agent, Runner
from agents.mcp import MCPServerStdio

async with MCPServerStdio(
    params={"command": "uvx", "args": ["keboola_mcp_server"]}
) as mcp:
    agent = Agent(
        name="Assistant",
        instructions="Use the Keboola tools to achieve the task",
        mcp_servers=[mcp],
    )
    result = await Runner.run(agent, "Load yesterday's CSV into Snowflake")

The SDK automatically calls list_tools() on the server, making every Keboola operation available to the model.

LangChain

LangChain provides an official MCP adapter, langchain-mcp-adapters. Point its MultiServerMCPClient at the Keboola MCP Server and load the tools into your agent:

from langchain_mcp_adapters.client import MultiServerMCPClient

client = MultiServerMCPClient(
    {
        "keboola": {
            "transport": "streamable_http",
            "url": "https://mcp.<YOUR_REGION>.keboola.com/mcp",
        }
    }
)
tools = await client.get_tools()
# pass `tools` to your LangChain / LangGraph agent

Other frameworks

CrewAI – Use the native MCP support via MCPServerAdapter from crewai-tools[mcp] to expose the Keboola MCP Server’s tools to your crew.

Building your own MCP client

If you are developing your own MCP client or integrating MCP capabilities into a custom application, you can connect to the Keboola MCP Server. The server supports standard MCP communication protocols.

For detailed instructions and SDKs for building your own MCP client, refer to the official Model Context Protocol documentation for client developers. Supported transports are stdio and Streamable HTTP. For more details on the Keboola MCP server, including how it can be run and configured for custom client integration, refer to its GitHub repository.

Advanced Setup Options

These methods are for developers or specific use cases (e.g., testing, contributing to the MCP server). Prefer a terminal or want to give an agent sandboxed, multi-project control? See the kbagent CLI — it can also call MCP tools via kbagent tool. For dev environments or contributing to the MCP Server, check out the MCP GitHub repo.

Support and Feedback

Need help or want to contribute? Open an issue on GitHub to report bugs, request features, or suggest improvements. We’d love your ideas, fixes, and feedback to make MCP even better.