claude-pytorch-treehugger

What is this MCP

This MCP provides a Python library and server for interacting with PyTorch's CI/CD data through a HUD API, offering access to workflows, job logs, and analytics with ClickHouse integration.

How to use this MCP

Install via pip (pip install git+https://github.com/izaitsevfb/claude-pytorch-treehugger.git) and start the MCP server with python -m pytorch_hud. Use CLI commands like claude mcp add hud pytorch-hud to connect.

What this MCP can be used for

Analyzing PyTorch CI/CD pipelines, extracting log patterns, monitoring test results, tracking resource utilization, and performing efficient log analysis for large-scale CI systems.