Exposes real-time NVIDIA GPU metrics like utilization, memory, temperature, and power to AI agents via the Model Context Protocol.
Also known as pmady/gpu-mcp-server, gpu-mcp-server
- Stars
- 15
- Tools
- 4
- Downloads, 30 days
- —
- Latest version
- 0.1.0
- Last commit
- Nov 14, 2026
- License
- Apache-2.0
What is GPU MCP?
GPU exposes real-time NVIDIA GPU metrics like utilization, memory, temperature, and power to AI agents via the Model Context Protocol. It exposes 4 tools, such as list_gpus, get_gpu_metrics, get_gpu_processes.
Summary generated from the project's README
About GPU MCP server
Monitoring NVIDIA hardware performance from within an AI agent is what this server enables. It provides real-time access to GPU metrics such as utilization, memory usage, temperature, and power draw. Developers can connect this server to compatible clients like Claude, Goose, Cursor, or Windsurf using a local stdio connection.
This tool functions as a local process that communicates directly with NVML via cgo, removing the need for complex metric pipelines like Prometheus or dcgm-exporter. It supports Multi-Instance GPU (MIG) configurations, where MIG instances appear as distinct devices while still reporting shared metrics from the parent hardware.
Users can interact with the hardware through specific tools including list_gpus, get_gpu_metrics, get_gpu_processes, and gpu_summary. The server can be deployed as a compiled Go binary or via Docker containers, provided the host machine has the necessary NVIDIA drivers and Container Toolkit installed.
What you can do with it
- 1Check hardware utilizationAsk your AI agent to list all available GPUs and their current memory and utilization status.
- 2Identify resource-heavy processesRequest a breakdown of which specific process IDs are currently using the GPU.
- 3Get aggregate system statsAsk for a summary of the total memory and average utilization across all connected devices.
Written from the project's README. Every use case is checked against a quote from it.
How to install GPU MCP
Formats verified 10 Nov 2026Add to CursorAdd to GooseInstall GPU MCP in Claude Code
$ claude mcp add gpu-mcp-server -- docker run -i --rm ghcr.io/pmady/gpu-mcp-server:v0.1.0Install GPU MCP in Cursor
~/.cursor/mcp.json
{
"mcpServers": {
"gpu-mcp-server": {
"command": "docker",
"args": [
"run",
"-i",
"--rm",
"ghcr.io/pmady/gpu-mcp-server:v0.1.0"
]
}
}
}Install GPU MCP in VS Code
.vscode/mcp.json
{
"servers": {
"gpu-mcp-server": {
"type": "stdio",
"command": "docker",
"args": [
"run",
"-i",
"--rm",
"ghcr.io/pmady/gpu-mcp-server:v0.1.0"
]
}
}
}Install GPU MCP in VS Code CLI
$ code --add-mcp '{"name":"gpu-mcp-server","type":"stdio","command":"docker","args":["run","-i","--rm","ghcr.io/pmady/gpu-mcp-server:v0.1.0"]}'Install GPU MCP in Claude Desktop
claude_desktop_config.json
{
"mcpServers": {
"gpu-mcp-server": {
"command": "docker",
"args": [
"run",
"-i",
"--rm",
"ghcr.io/pmady/gpu-mcp-server:v0.1.0"
]
}
}
}Install GPU MCP in Codex
$ codex mcp add gpu-mcp-server -- docker run -i --rm ghcr.io/pmady/gpu-mcp-server:v0.1.0Install GPU MCP in Codex config.toml
~/.codex/config.toml
[mcp_servers.gpu-mcp-server] command = "docker" args = ["run", "-i", "--rm", "ghcr.io/pmady/gpu-mcp-server:v0.1.0"]
Install GPU MCP in Gemini CLI
$ gemini mcp add -s user gpu-mcp-server docker -- run -i --rm ghcr.io/pmady/gpu-mcp-server:v0.1.0Install GPU MCP in Gemini settings.json
~/.gemini/settings.json
{
"mcpServers": {
"gpu-mcp-server": {
"command": "docker",
"args": [
"run",
"-i",
"--rm",
"ghcr.io/pmady/gpu-mcp-server:v0.1.0"
]
}
}
}GPU MCP tools 4
Detected from the READMElist_gpusget_gpu_metricsget_gpu_processesgpu_summary
| list_gpus | — | |
| get_gpu_metrics | — | |
| get_gpu_processes | — | |
| gpu_summary | — |
Automated checks
Ran Nov 12, 2026Full report →Passed our automated checks. That is not a security audit: read the code and give the server the narrowest access it needs. How checks work
- Source repository linked and reachablePass
- Open source license (OSI)Pass
- Commit or release in the last yearPass
- Package name not close to a popular packageNot checked
- Remote endpoint uses HTTPSNot checked
- Destructive tools are labelledPass
- No hidden instructions to the modelPass
- No critical advisory in the package or its dependenciesNot checked
README
View on GitHubgpu-mcp-server
Note: the OpenSSF Best Practices questionnaire is in progress. Once the project entry is registered at https://www.bestpractices.dev/en, swap the static badge above for the live one:
[](https://www.bestpractices.dev/projects/<ID>)
An MCP server that exposes NVIDIA GPU metrics as tools. Any MCP-compatible AI agent (Claude, Goose, Cursor, etc.) can query real-time GPU utilization, memory, temperature, power, PCIe and NVLink throughput no Prometheus or dcgm-exporter required.
Built on the official Go MCP SDK and NVIDIA go-nvml.
Tools
| Tool | Description |
|---|---|
list_gpus |
List all GPUs with utilization and memory info |
get_gpu_metrics |
Detailed metrics for a GPU by index or UUID |
get_gpu_processes |
PID-level GPU process attribution |
gpu_summary |
Aggregate stats across all devices |
All tools support MIG (Multi-Instance GPU) - MIG instances appear as separate devices with their parent GPU's shared metrics (temperature, power, PCIe).
Sample output
Each tool returns structured JSON. The examples below show the shape of the data an agent receives from a node with two NVIDIA A100 GPUs.
list_gpus:
{
"count": 2,
"devices": [
{
"index": 0,
"uuid": "GPU-aaaa-1111",
"name": "NVIDIA A100-SXM4-80GB",
"gpu_utilization_percent": 85,
"memory_used_mib": 57344,
"memory_total_mib": 81920
},
{
"index": 1,
"uuid": "GPU-bbbb-2222",
"name": "NVIDIA A100-SXM4-80GB",
"gpu_utilization_percent": 20,
"memory_used_mib": 12288,
"memory_total_mib": 81920
}
]
}
get_gpu_metrics (with {"index": 0} or {"uuid": "GPU-aaaa-1111"}):
{
"index": 0,
"uuid": "GPU-aaaa-1111",
"name": "NVIDIA A100-SXM4-80GB",
"gpu_utilization_percent": 85,
"memory_utilization_percent": 70,
"memory_used_mib": 57344,
"memory_total_mib": 81920,
"temperature_celsius": 72,
"power_draw_watts": 300,
"power_limit_watts": 400,
"pcie_tx_kbps": 0,
"pcie_rx_kbps": 0,
"nvlink_tx_mbps": 0,
"nvlink_rx_mbps": 0
}
gpu_summary:
{
"device_count": 2,
"avg_gpu_utilization": 52.5,
"avg_memory_utilization": 42.5,
"total_memory_used_mib": 69632,
"total_memory_total_mib": 163840,
"max_temperature_celsius": 72,
"total_power_draw_watts": 375
}
MIG instances add is_mig, parent_gpu, and mig_profile fields to the
get_gpu_metrics and list_gpus payloads.
Quick start
# build (requires CGO + NVML headers on Linux)
make build
# run the server communicates over stdio
./gpu-mcp-server
Claude Desktop
Add to claude_desktop_config.json:
{
"mcpServers": {
"gpu": {
"command": "/path/to/gpu-mcp-server"
}
}
}
Goose
extensions:
gpu-metrics:
type: stdio
cmd: /path/to/gpu-mcp-server
Cursor
Add to .cursor/mcp.json for a project, or ~/.cursor/mcp.json for all
projects:
{
"mcpServers": {
"gpu": {
"type": "stdio",
"command": "/path/to/gpu-mcp-server"
}
}
}
Windsurf
Add to ~/.codeium/windsurf/mcp_config.json:
{
"mcpServers": {
"gpu": {
"command": "/path/to/gpu-mcp-server"
}
}
}
Build
Requires Go 1.23+, CGO, and NVIDIA drivers on the target machine.
make build # compile binary
make test # run tests (no GPU needed uses mock)
make lint # golangci-lint
make docker # container image
Tests use a mock collector, so they run anywhere no GPU hardware required.
Docker
Prebuilt multi-arch images (linux/amd64, linux/arm64) are published to GHCR on every release.
docker pull ghcr.io/pmady/gpu-mcp-server:latest
docker run --rm -i --gpus all ghcr.io/pmady/gpu-mcp-server:latest
The host needs the NVIDIA Container Toolkit
installed for --gpus all to work. The server speaks MCP over stdio, so the
-i flag is required — don't drop it.
{
"mcpServers": {
"gpu": {
"command": "docker",
"args": ["run", "--rm", "-i", "--gpus", "all", "ghcr.io/pmady/gpu-mcp-server:latest"]
}
}
}
Pin a specific version via tag instead of :latest, e.g. ghcr.io/pmady/gpu-mcp-server:v0.1.0.
Architecture
Agent (Claude/Goose) ─── MCP (stdio) ──→ gpu-mcp-server ──→ NVML ──→ GPU
│
Tools:
• list_gpus
• get_gpu_metrics
• gpu_summary
The server runs as a local process alongside the agent. It calls NVML directly through cgo — no sidecar, no network hops, no metric pipeline to configure.
Project info
- License: Apache 2.0
- Language: Go
- AAIF project alignment: MCP
- Related: keda-gpu-scaler (GPU autoscaling for Kubernetes)
- Whitepaper: GPU-Aware Autoscaling in Cloud Native AI Infrastructure — CNCF TAG Infrastructure initiative (TOC #2188)
Citing
If you use gpu-mcp-server in your research or writing, please cite it via its DOI. Full citation metadata is in CITATION.cff.
@software{madduri_gpu_mcp_server,
author = {Madduri, Pavan},
title = {gpu-mcp-server: NVIDIA GPU metrics for AI agents over the Model Context Protocol},
year = {2026},
publisher = {Zenodo},
doi = {10.5281/zenodo.22866670},
url = {https://doi.org/10.5281/zenodo.22866670}
}
Roadmap
See ROADMAP.md for the 12-month public roadmap.
Contributing
See CONTRIBUTING.md for how to get involved.
Contributors
Thanks to all our contributors! Add yourself via PR.
Governance
This project follows Linux Foundation Minimum Viable Governance.
Documentation
- Full documentation - hosted on Read the Docs
- ROADMAP.md - public roadmap
- GOVERNANCE.md - decision-making process
- DEPENDENCIES.md - external dependencies and licenses
- SECURITY.md - vulnerability reporting
- AGENTS.md - instructions for AI agents working on this repo
- CODE_OF_CONDUCT.md - community standards
Star History
GPU MCP questions
How do I add GPU MCP to Claude Code?
Run claude mcp add gpu-mcp-server -- docker run -i --rm ghcr.io/pmady/gpu-mcp-server:v0.1.0 in your terminal. Add --scope user to make it available in every project.
How do I add GPU MCP to Cursor?
Open ~/.cursor/mcp.json and add the server under mcpServers, using the Cursor config shown in the install section above.
What tools does GPU MCP provide?
It exposes 4 tools, including list_gpus, get_gpu_metrics, get_gpu_processes, gpu_summary.
Does this require Prometheus or dcgm-exporter?
No, the server queries NVIDIA GPU metrics directly without needing a Prometheus or dcgm-exporter setup.
Does it support Multi-Instance GPU (MIG)?
Yes, all tools support MIG, where instances appear as separate devices with shared parent metrics like temperature and power.
Can I run this using Docker?
Yes, prebuilt multi-arch Docker images are available on GHCR, provided the host has the NVIDIA Container Toolkit installed.