denpex-mcp
OfficialProvides diagnostics for NVIDIA GPU and Xid-related failures, including routing Xid faults to appropriate node actions, interpreting nvidia-smi telemetry, and referencing NVIDIA Xid error information.
Provides diagnostics for PyTorch/ML training crashes and checks known-bad CUDA/PyTorch/NCCL/driver version combinations as part of failure analysis.
This server does not currently have a deployable release on Glama. Follow the repository's setup instructions to install and configure it for your MCP client, including any required credentials and local dependencies.
Denpex MCP server
Log files up to 100 MB are read in bounded chunks and reduced to diagnostic evidence before cloud submission. Binary files and invalid UTF-8 are rejected. Keep the original log: evidence selection can omit context.
Diagnose GPU/ML training failures. CUDA OOM, NCCL timeouts, Xid faults, NaN loss,
checkpoint corruption, without leaving your AI tool. Ask Claude Desktop, Cursor,
or Claude Code to "diagnose the crash in ./logs/train.err" and Denpex reads the log
straight off disk and hands back the root cause and the fix.
🩺 Root cause + escalating fix (primary → fallback → emergency), verification command, and prevention.
🧠Find the culprit rank in a distributed/NCCL hang, and route GPU Xid faults to the right node action, offline, instantly.
📚 Bundled failure encyclopedia (500+ documented classes) searchable by error code or symptom.
🔒 Zero-egress mode (
DENPEX_LOCAL=1): diagnose entirely on-host, nothing leaves the machine. Plus client-side PII scrubbing even in cloud mode.
Quick start (60 seconds)
1. Get an API key (optional but recommended)
Sign in at denpex.com/dashboard → API Keys and create a key (dpx_…).
The key unlocks authenticated cloud diagnosis, your diagnosis history, quota, and fleet
incidents. Without a key, diagnose_crash and ask_followup default to running the
canonical deterministic engine entirely locally (unlimited free local runs). You can escalate
a specific diagnosis to the Denpex cloud without a key by passing cloud: true, which uses
the frictionless anonymous engine (requires an API key or browser check if anonymous quota
is exhausted). Set DENPEX_LOCAL=1 explicitly for permanent zero-egress.
2. Add it to your client
Edit claude_desktop_config.json
(macOS: ~/Library/Application Support/Claude/, Windows: %APPDATA%\Claude\):
{
"mcpServers": {
"denpex": {
"command": "npx",
"args": ["-y", "denpex-mcp"],
"env": { "DENPEX_API_KEY": "dpx_your_key_here" }
}
}
}Restart Claude Desktop.
Create .cursor/mcp.json in your project (or ~/.cursor/mcp.json globally):
{
"mcpServers": {
"denpex": {
"command": "npx",
"args": ["-y", "denpex-mcp"],
"env": { "DENPEX_API_KEY": "dpx_your_key_here" }
}
}
}claude mcp add denpex -e DENPEX_API_KEY=dpx_your_key_here -- npx -y denpex-mcp3. Use it
"My training job crashed, diagnose the log at
./logs/rank0.err." "What does Xid 79 mean?" "This 64-GPU run is hung. Which rank failed first?" (→analyze_cascade)
Or pick the diagnose-training-crash prompt your client surfaces.
Related MCP server: Inference AIops
Offline / air-gapped mode
Diagnose entirely on-host, no network, no logs leaving the cluster. Ideal for VPCs,
secure data centers, and privacy-sensitive labs. Add DENPEX_LOCAL=1:
{ "mcpServers": { "denpex": {
"command": "npx", "args": ["-y", "denpex-mcp"],
"env": { "DENPEX_LOCAL": "1" }
} } }The package bundles the same canonical deterministic diagnosis pipeline used by the hosted wrapper: 10,700+ patterns, causal detectors and contradiction gates, Xid routing, cross-rank cascade localization, and the 500+ class encyclopedia. It returns the initiating rank/node when the evidence identifies one. Only the hosted LLM fallback and live-research enrichment are unavailable offline.
Tools
Tool | What it does | Needs |
| Root cause + escalating fix from | local-first by default without a key; cloud with key or |
| Search the bundled encyclopedia (500+ classes) by code/keyword/symptom. | works offline |
| Find the culprit rank in a distributed hang; route Xid faults. | works offline |
| Report whether a fix worked. This is what makes the next diagnosis better. | API key |
| Your plan and remaining diagnoses this month. | API key |
| Recall your recent diagnoses and their fixes. | API key |
| Conversational follow-up on a diagnosis. | local text-extraction without a key; generative chat with Team plan / trial |
| Flag known-bad CUDA/PyTorch/NCCL/driver combos. | Team plan / trial |
| "It worked yesterday", isolate what changed between a good run and a bad one. | Team plan / trial |
| Diff torch/CUDA/NCCL/driver versions and env vars between two runs. | Team plan / trial |
| Find the newest checkpoint that is actually safe to resume from. | Team plan / trial |
| "Is this cluster safe to launch on?", blockers before you burn GPU-hours. | Scale plan |
| Score every monitored node, worst first. | Scale plan |
| Nodes trending toward a failure that has not happened yet. | Scale plan |
| Incidents Denpex's monitor caught & escalated on your fleet. | Scale plan |
Gated tools degrade gracefully, they tell you exactly which plan unlocks them and which
tools work on your current plan. The failure encyclopedia is also exposed as resources at
denpex://failures/{slug}.
Structured output. Every tool above returns structuredContent alongside its
markdown, validated against a declared output schema, so an agent branches on
confidenceBand === "low" or iterates rankedHypotheses instead of regexing prose.
Two fields deserve attention when you write the agent loop:
confidence: nulldoes not mean zero. It means the engine measured this band's historical accuracy and declined to state a number. Treat the verdict as a lead. Do not substitute the raw score.unfalsified: trueidentifies only the explicitly enabled reduced compatibility fallback, where the canonical contradiction graph did not run. Normal cloud andDENPEX_LOCAL=1diagnosis both use the canonical engine and set it to false.
Environment variables
Variable | Default | Purpose |
| , | Your |
|
| Override the API base (self-host / staging). |
|
|
|
|
|
|
Develop
npm install
npm run build # regenerates the canonical engine + corpus, then bundles dist/index.js
npm test # 111 unit + protocol-integration tests
npm run inspector # interactive protocol inspector (MCP Inspector)The offline corpus (src/data/denpex_patterns.json) and canonical engine
(src/offline/canonical-engine.mjs) are generated from the repo's single source of truth
during npm run build, never hand-edit either artifact.
Official Denpex Ecosystem & Documentation
Resource | Description | Link |
Denpex Platform | Deterministic GPU and ML training failure diagnostics | |
Documentation Hub | Quickstart guides, architecture, and operational reference | |
Python Agent & CLI | On-host monitoring, evidence capture, and zero-egress CLI | |
REST API Reference | Programmatic | |
VS Code Extension | Offline GPU crash diagnostics in VS Code | |
Open VSX Extension | Extension for Cursor, Windsurf, VSCodium, and Gitpod | |
Failure Encyclopedia | 500+ documented CUDA, NCCL, Xid, Slurm, and vLLM failures | |
Free GPU Tools | VRAM, NCCL, DeepSpeed ZeRO, and Slurm calculators | |
NVIDIA Xid Reference | Hardware vs software Xid fault routing table | |
Release Changelog | Platform, agent, SDK, and extension release notes |
Free Web Calculators & Xid Reference
63-Model GQA/MLA LLM VRAM Calculator: denpex.com/tools/vram-calculator
GRPO / RLHF VRAM Calculator (verl / vLLM): denpex.com/tools/grpo-rl-vram-calculator
Zero-Upload Browser GPU Log Analyzer: denpex.com/tools/gpu-log-analyzer
NVIDIA Xid Error Directory: denpex.com/xid
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
This server cannot be deployed
Maintenance
Related MCP Connectors
Diagnose why an AI agent failed and get the verified fix instantly. Free, no token.
AI SRE: investigate production incidents across your observability stack.
Diagnose AI workflows for failure, security, and handoff risks — RED/AMBER/GREEN per node.
Finds what's broken in your product, explains why, and fixes it from your agent.
Related MCP Servers
- AlicenseBqualityDmaintenanceAI-powered Kubernetes diagnostics that analyzes pod crashes, logs, and cluster health to provide root cause analysis and actionable solutions for common issues like CrashLoopBackOff, OOM kills, and connection errors.811 npm1MIT

Inference AIopsofficial
AlicenseBqualityBmaintenanceEnables governance-grade AIops for GPU inference clusters with root-cause analysis, metrics, and policy-governed operations for vLLM and Ray.39MIT- FlicenseNot gradedqualityAmaintenanceEnables AI assistants to run GPU/CUDA environment diagnostics, check hardware and software compatibility, get safe install commands, and validate Dockerfiles for GPU issues through natural-language requests.-
- FlicenseNot gradedqualityAmaintenanceEnables AI assistants to invoke GPU/CUDA diagnostics for checking driver, toolkit, cuDNN, Python-library compatibility, model fit, Dockerfile GPU issues, and safe install commands.-