Sign inGet started
Repository · npm

neurarch/neurarch-mcp

MCP server that tells Claude Code, Cursor, Claude Desktop, VS Code and other MCP-aware agents what a training job needs before it takes a GPU: size_training_job returns the calibrated training peak at your batch, the smallest card that holds it, max batch

Ownerneurarch
License (as declared)MIT
Stars0
Forks0
Main languagejavascript
Created20 October 2026
Last updated at source20 October 2026
npmjavascriptarchitecture-searchclaudeclaude-desktopcode-reviewcodexcopilotcursorhuggingfacelintmachine-learningmcpmodel-context-protocolneural-networkneurarchpytorchpytorch-lintstatic-analysisvscodewindsurf
Embed this entry
<iframe src="https://silytics.si/embed/item/4185d669-ac92-41e5-af1b-ff2a5a68361c" width="420" height="190" style="border:0" loading="lazy" title="Silytics entry"></iframe>

Free, no tracking. More about feeds and embeds.

Stars, last 6 days

908 now (+528 since 2026-10-15). Counts as reported by the source, snapshotted daily by Silytics.

Recent changes

  • 18 Oct 2026 Description: MCP server that gives Claude Code, Cursor, Claude Desktop, VS Code and other MCP-aware agents structural awareness of a neural network. Point it at a PyTorch .py file, a Neurarch graph, or a Hugging Face repo and the agent can read layers, → MCP server that tells Claude Code, Cursor, Claude Desktop, VS Code and other MCP-aware agents what a training job needs before it takes a GPU: size_training_job returns the calibrated training peak at your batch, the smallest card that hold

Changes detected by Silytics between refreshes of the source.

Silytics shows public metadata from npm and links to the original. Information may be incomplete or out of date. Last refreshed 20 October 2026. Licenses and terms of the original apply to the work itself. Not financial, legal or professional advice.