Keboola MCP Server

🚀 Keboola MCP Server Manual

The Keboola MCP (Management Console) Server is a tool for interacting with Keboola. Here are the main functions of this server:

✨ Features

Storage Tools

• retrieve_buckets - List all storage buckets in the project.
• get_bucket_detail - Get detailed information about a specific storage bucket.
• retrieve_bucket_tables - List all tables in a specific storage bucket.
• get_table_detail - Get detailed information about a specific table.
• update_bucket_description - Update the description of a storage bucket.
• update_table_description - Update the description of a table.

SQL Tools

• query_table - Execute SQL queries in the workspace.
• get_sql_dialect - Get the SQL dialect (Snowflake or BigQuery) used in the workspace.

Component Tools

• retrieve_components - List all available components and their configurations.
• retrieve_transformations - List transformation configurations.
• get_component_details - Get detailed information about a specific component.
• create_sql_transformation - Create a new SQL transformation configuration.

Job Tools

• retrieve_jobs - List all jobs in the project.
• get_job_detail - Get detailed information about a specific job.
• start_job - Start a job with a new component configuration.

Documentation Tools

• docs_query - Query documentation and help information.

📦 Installation and Configuration

Install the MCP Server

1. Clone the Keboola MCP Server code repository to your local machine.
2. Activate the virtual environment and install the dependencies:

   python -m pip install -r requirements.txt
   

Configure the MCP Server

• Set the following environment variables:
  • KBC_API_URL: The URL of the Keboola API (e.g., https://connection.YOUR_REGION.keboola.com).
  • KBC_STORAGE_TOKEN: Your storage token.
  • KBC_WORKSPACE_SCHEMA: Your workspace schema (Snowflake or BigQuery).

💻 Usage Examples

Start the MCP Server

python -m keboola_mcp_server.cli --transport <transport-type> --api-url https://connection.YOUR_REGION.keboola.com

Supported transport types include:

• http: Access via HTTP endpoints.
• stdio: Interact via standard input and output.
• sse: Interact via Server-Sent Events.

Use MCP Tools

Here are some usage examples of common tools:

Storage Tool Examples

# Retrieve all storage buckets
python -m keboola_mcp_server.cli retrieve-buckets --transport stdio

# Get detailed information about a specific storage bucket
python -m keboola_mcp_server.cli get-bucket-detail <bucket-id> --transport stdio

SQL Tool Examples

# Execute an SQL query
python -m keboola_mcp_server.cli query-table <table-id> "<your-sql-query>" --transport stdio

# Get the SQL dialect
python -m keboola_mcp_server.cli get-sql-dialect --transport stdio

Component Tool Examples

# Retrieve all components
python -m keboola_mcp_server.cli retrieve-components --transport stdio

# Create a new SQL transformation configuration
python -m keboola_mcp_server.cli create-sql-transformation <component-id> "<transformation-query>" --transport stdio

📄 Important Notes

⚠️ Important Note

Before using the MCP Server in a production environment, ensure that all environment variables are correctly set.

💡 Usage Tip

Regularly update dependencies to get the latest features and security patches.

🔧 Development and Testing

Run Tests

python -m pytest tests/

Submit Code

Follow the standard Git process for submissions:

1. Create new features or fix issues.
2. Commit changes to your local repository.
3. Push to the remote repository and create a Pull Request.

🛠️ Troubleshooting

If you encounter any issues, refer to the Keboola Documentation or contact the support team.

Through the above steps, you can successfully install, configure, and use the Keboola MCP Server, and develop and expand it as needed.