Installation & Setup

Installation

Make sure you have Python 3.10+ installed and use the following command to install the client. It’s recommended to use a virtual environment.

This installs the CLI that you and your coding agent can use to run tasks, actions and agents with Ellf and manage your cluster. It also sets up everything you need for your local development environment. Next, you can authenticate and connect Ellf to your coding assistant.

Claude Code Anthropic

Logging in with the --claude flag downloads and sets up everything that’s needed for the Claude Code integration. The custom Ellf skills for the different modules should then show up in your slash commands.

/ellf-annotate    Define labels, components and annotate data
/ellf-ask         Ask specific questions and query the knowledge base
/ellf-monitor     Monitor jobs, training progress and cluster health
/ellf-ops         Navigate the platform and CLI to perform actions
/ellf-patterns    Interactively develop patterns and rule-based logic
/ellf-prodigy     Implement workflows for the Prodigy annotation tool
/ellf-project     Plan the project and break down business problem
/ellf-support     Get human support and feedback from our NLP experts
/ellf-todo        Pick up a coding task from the in-app assistant
/ellf-train       Train and export custom models and components

Next, add the Ellf MCP server via the claude mcp add command:

Codex OpenAI

Cursor

Copliot GitHub

Configuration