Rescale MCP Server

Overview

The Rescale MCP Server is a local application that bridges Large Language Models (LLMs) and the Rescale platform. It solves the challenge of LLMs being unable to directly or securely interact with web APIs by implementing the open-source Model Context Protocol (MCP) on top of Rescale Platform APIs.

Running on your local machine, the server translates natural language commands from an MCP client (like Claude Desktop) into authenticated Rescale API actions, abstracting away the integration complexity. This is the same technology that powers the Rescale Assistant.

How it Works

The workflow is simple and secure. When you issue a command in an AI client, it’s sent to the local MCP server. The server then authenticates your request using your Rescale API token, translates the command into the appropriate API call, executes it, and returns the result to your client. This process allows you to interact with the Rescale platform using your local AI tools.

data documentation image

Getting Started

The Rescale MCP Server already powers the Rescale Assistant, using a managed version to process your requests seamlessly.

For developers wanting to connect their local clients like Claude Desktop, the Rescale MCP server can be installed from our GitHub repository and configured with your client path and Rescale API token. For complete instructions and source code, visit the official repository.

Core Capabilities

By connecting local clients to Rescale, you can perform a wide range of actions using natural language.

Interact with Natural Language

Replace complex scripting and API calls with simple, conversational phrases to make your workflow more intuitive.

  • Example: Ask to “submit a job using input_file.zip on the Onyx core type” or “show me the status of job XYZ.”

Manage Jobs and Workstations

Control the entire lifecycle of your Rescale compute jobs and workstations directly from your client.

  • Create and Submit: Configure and launch new jobs or workstations.
  • Monitor and Control: Track the status of active jobs, view logs, and perform actions like stopping or terminating workloads.

Transfer Files

Streamline data transfer between your local machine and the Rescale cloud platform.

  • Upload and Download: Send input files for your simulations and retrieve results directly to your desktop.
  • Manage Data: Organize files and folders associated with your Rescale jobs.

For a complete list of available tools and their usage, see our MCP documentation.

×