Hugging Face's logo

Spaces:
tuntun
/
MLflow_mcp
Sleeping

App Files Community
MLflow_mcp / README.md
tuntun's picture
tuntun
update
b0f3d83
preview code
|
raw
history blame contribute delete
1.29 kB

A newer version of the Gradio SDK is available: 5.44.1

Upgrade
metadata
title: MLflow Mcp
emoji: 👁
colorFrom: indigo
colorTo: purple
sdk: gradio
sdk_version: 5.33.0
app_file: app.py
pinned: false
license: mit
short_description: MCP service for MLFlow
tags:
  - mcp-server-track

Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference

Gradio-Based MCP Service Design

1. Overview

This is a Gradio application that exposes a set of MLflow "tools" (functions) as JSON/HTTP endpoints. An LLM (e.g., Claude or another agent) can invoke these endpoints to manage experiments, runs, and models. A minimal form-per-function UI also allows humans to issue the same commands via a Hugging Face Space.

Core functions (MCP tools):

  1. Set MLflow tracking URI
  2. Get system info (including tracking URI, registry URI, MLflow version, etc.)
  3. List all experiments
  4. Create a new experiment
  5. Register a model artifact
  6. Search runs by filter
  7. List all registered models
  8. Get detailed info about a specific registered model

2. Demos

Due to time and resource limitation, this demo shows a MLFLow server run locally and also the MCP server. Please input a remote MLFlow server if you would like to use the config in HuggingFace space. Demo video in youtube