A lightweight Model Context Protocol (MCP) server that enables AI assistants to interact with ClearML experiments, models, and projects for comprehensive ML experiment context analysis.

Installation

PyPI

pip install clearml-mcp

uvx (no installation)

uvx clearml-mcp

From source

git clone https://github.com/prassanna-ravishankar/clearml-mcp.git
cd clearml-mcp
uv sync
uv run python -m clearml_mcp.clearml_mcp

Configuration

Claude Desktop

{
  "mcpServers": {
    "clearml": {
      "command": "uvx",
      "args": ["clearml-mcp"]
    }
  }
}

Cursor

{
  "mcp.servers": {
    "clearml": {
      "command": "uvx",
      "args": ["clearml-mcp"]
    }
  }
}

Continue

{
  "mcpServers": {
    "clearml": {
      "command": "uvx",
      "args": ["clearml-mcp"]
    }
  }
}

Cody

{
  "cody.experimental.mcp": {
    "servers": {
      "clearml": {
        "command": "uvx",
        "args": ["clearml-mcp"]
      }
    }
  }
}

Available Tools

Tool	Description
get_task_info	Get detailed task information, parameters, and status
list_tasks	List tasks with advanced filtering (project, status, tags, user)
get_task_parameters	Get hyperparameters and configuration
get_task_metrics	Access training metrics, scalars, and plots
get_task_artifacts	Get artifacts, model files, and outputs
get_model_info	Get model metadata and configuration details
list_models	View available models with filtering
get_model_artifacts	Access model files and download URLs
list_projects	Discover available ClearML projects
get_project_stats	Get project statistics and task summary
find_project_by_pattern	Find projects matching a name pattern
find_experiment_in_project	Find specific experiments in projects
compare_tasks	Compare multiple tasks by specific metrics
search_tasks	Advanced search by name, tags, comments, and more

Features

• Experiment Discovery: Search and analyze ML experiments in projects
• Performance Analysis: Compare model metrics and training progress
• Real-time Metrics: Access training scalars, validation curves, and convergence analysis
• Smart Search: Filter tasks by name, tags, status, and custom queries
• Artifact Management: Retrieve model files, datasets, and experiment results
• Cross-platform: Works with all major AI assistants and code editors

Usage Examples

Show me the latest experiments in the 'computer-vision' project

Compare accuracy metrics between tasks task-123 and task-456

What are the hyperparameters of the best-performing model?

Find all failed experiments from last week

Get the training curves for my latest BERT fine-tuning

Notes

Requires a ClearML account with valid API credentials in ~/.clearml/clearml.conf. Provides 14 comprehensive tools for ML experiment analysis. Compatible with Zed Editor, OpenHands, Roo-Cline, and any MCP-enabled applications.