A comprehensive MCP server that provides machine learning model training, fine-tuning, and experiment management functions, supporting multi-backend training, cloud GPU deployment, and cost estimation.

An instant SRE diagnostic agent for NVIDIA GPU hardware diagnosis in Kubernetes clusters, providing real-time GPU hardware detection and fault troubleshooting functions through the MCP protocol.

VkRunner is a Vulkan shader testing tool based on Piglit's shader_runner, aiming to support test scripts as similar as possible to Piglit's shader_test format. It supports writing shaders in GLSL or SPIR-V and provides a rich set of test commands to verify shader behavior, including drawing, computing, and probe detection functions.

The Hyperbolic GPU MCP Server is a tool for interacting with the Hyperbolic GPU cloud, allowing agents and LLMs to view, rent GPUs, and run GPU-accelerated workloads via SSH connections.

The Docker deployment solution for Kyutai TTS provides a one-click startup web interface, REST API, and MCP tool support, supporting GPU acceleration and multilingual interfaces.

This project integrates Graphistry's GPU-accelerated graph visualization platform with the Model Control Protocol (MCP), providing advanced graph analysis capabilities for AI assistants and large language models, and supporting multiple data formats and network analysis functions.

The Hyperbolic GPU MCP server is a Node.js - based tool that allows users to manage and rent GPU resources on the Hyperbolic cloud platform via an API, including functions such as viewing available GPUs, renting instances, SSH connection, and running GPU workloads.

An all-in-one Docker image for the QuantConnect Lean algorithmic trading engine, supporting automatic GPU selection, a modern web interface, REST API, and MCP protocol integration

The MCP Speaker Diarization and Recognition System is a complete solution integrating GPU-accelerated speaker separation, speech recognition, emotion detection, and a web interface. It combines the speaker separation of pyannote.audio with the faster-whisper transcription technology, supporting persistent speaker recognition (one-time registration, permanent recognition), dual-detector emotion analysis (combining general AI and personalized voiceprints), real-time stream processing, REST API, and MCP server, and is designed for AI intelligent agent integration and hobby projects.

An MCP server for allocating resources and executing Python code on Google Colab's GPU runtime (T4/L4), enabling AI assistants to remotely run GPU-accelerated computing tasks.