cacheout-mcpofficial
A
licenseA
qualityC
maintenanceMCP server for macOS disk cache management. Enables AI agents to free disk space on demand by scanning and clearing developer caches.
13
22 PyPI
1
Sustainable Use v1.0
Best PyTorch MCP Servers
PyTorch is an open source machine learning framework based on the Torch library, used for applications such as computer vision and natural language processing.
Why this server?
Allows cleaning of PyTorch model cache (torch hub).
Why this server?
Supports deployment of GPU instances using PyTorch/CUDA images as part of the Novita AI platform resource management capabilities.
Novita MCP Serverofficial
AlicenseCqualityDmaintenanceAn MCP server that enables seamless management of Novita AI platform resources, currently supporting GPU instance operations (list, create, start, stop, etc.) through compatible clients like Claude Desktop and Cursor.2053 npm12MITWhy this server?
Provides diagnostic capabilities specifically for PyTorch workloads, identifying bottlenecks in the forward pass, data movement, and the caching allocator.
AlicenseAqualityAmaintenanceeBPF-based GPU causal observability agent with MCP server. Traces CUDA Runtime and Driver APIs via kernel uprobes and host events via tracepoints to build causal chains explaining GPU latency. 7 tools: get_check, get_trace_stats, get_causal_chains, get_stacks, run_demo, get_test_report, run_sql. Telegraphic compression reduces token usage ~60%. Supports stdio and HTTPS (TLS 1.3) transport.1184-Why this server?
Enables scanning of PyTorch checkpoint files (.pt, .pth) for malicious pickle members and other deserialization-based security threats.
AlicenseBqualityDmaintenanceModelSafetyMCP is an MCP server for scanning machine learning model artifacts for unsafe serialization, malicious model patterns, risky packaging, URL-based artifact scanning, and directory-level triage.61MITWhy this server?
Enables deployment of pods with PyTorch images, allowing users to specify PyTorch-based container environments.
RunPod MCP Serverofficial
AlicenseCqualityAmaintenanceThis Model Context Protocol server enables interaction with RunPod's REST API through Claude or other MCP-compatible clients, providing tools for managing pods, endpoints, templates, network volumes, and container registry authentications.366,573 npm86Apache 2.0Why this server?
Uses PyTorch as the underlying framework for model operations, with optimizations for faster training and reduced memory usage.
AlicenseAqualityDmaintenanceProvides tools for optimizing, fine-tuning, and deploying large language models with Unsloth, enabling 2x faster training with 80% less memory through model loading, fine-tuning, text generation, and model export capabilities.68Apache 2.0Why this server?
Fetches official PyTorch documentation, including tensor operations and neural network modules.
AlicenseAqualityDmaintenanceEnables managing internal documentation with search and add capabilities, and fetching official documentation from 13+ popular libraries via MCP tools.69MITWhy this server?
Supports PyTorch model format for Hailo-8 AI accelerator inference workloads.
AlicenseBqualityAmaintenanceEnables AI assistants to manage homelab infrastructure through automated service installation (Jellyfin, Pi-hole, Ollama, Home Assistant, Frigate NVR), VM operations, AI accelerator support (MemryX, Coral TPU, Hailo-8), and Terraform state management with SSH-based discovery and deployment.585MITWhy this server?
Integrates with a local PyTorch interpreter to profile and optimize kernels: traces a script's top CUDA kernels by device time via torch.profiler, runs scripts repeatedly to collect their timing line (mean/std/min), and runs baseline vs. candidate interleaved to report speedup and a verdict (faster, slower, inconclusive, incorrect, invalid). Also scaffolds targets from arbitrary PyTorch functions, using their fp32 result as the correctness reference.