mcp

The Model Context Protocol (MCP) Server is a core component of the Keboola Platform, designed to facilitate seamless data interactions and model management. Developers can utilize this server to efficiently manage and deploy machine learning models, ensuring that your data workflows are both robust and scalable.

Once connected to the MCP server, you can interact with it via HTTP requests to manage your machine learning models and data contexts. You can send commands to upload models, query model statuses, and retrieve contextual data. For optimal performance, leverage batch processing commands to handle large datasets efficiently.

Prerequisites

Ensure you have Node.js and npm installed on your system, as the MCP server relies on these tools for installation and operation.

Option A: Quick start with npx

You can quickly start using the MCP server without a global installation by running:

npx -y keboola/mcp-server

Option B: Global install alternative

To install the MCP server globally, use the following command:

npm install -g keboola/mcp-server

When configuring the MCP server, be mindful of the required environment variables, such as API keys and database connection strings, which are essential for proper operation. Additionally, watch for common issues related to network permissions and make sure your firewall settings allow traffic to and from the server.