shea256

shea256 / autofoundry

Public

A CLI tool that automates the provisioning of GPU's across cloud providers and the running of AI experiments across them

13
0
100% credibility
Found Mar 17, 2026 at 13 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
Python
AI Summary

Autofoundry is a tool for easily running machine learning experiment scripts in parallel across rented GPU computers from various providers, with automatic setup, live monitoring, and results reporting.

How It Works

1
🔍 Discover Autofoundry

You find this helpful tool that simplifies running AI training experiments on powerful rented computers.

2
📥 Set it up

You grab the tool and prepare it on your computer with a simple download and install.

3
🔗 Connect GPU services

You link a few GPU rental services by sharing your login details once, so everything works smoothly.

4
📝 Pick your experiment

You choose a simple script for your AI training and decide how many tests to run.

5
🚀 Launch the magic

You select the computer type like a super-fast H100, hit go, and it grabs the cheapest available rentals automatically.

6
👀 Watch live action

You see real-time updates as computers spin up, experiments run side-by-side, and outputs stream in.

✅ Get your insights

It wraps up with a clear summary of the best, average, and worst results, then shuts everything down safely.

Sign up to see the full architecture

5 more

Sign Up Free

Star Growth

See how this repo grew from 13 to 13 stars Sign Up Free
Repurpose This Repo

Repurpose is a Pro feature

Generate ready-to-use prompts for X threads, LinkedIn posts, blog posts, YouTube scripts, and more -- with full repo context baked in.

Unlock Repurpose
AI-Generated Review

What is autofoundry?

Autofoundry is a Python CLI tool that provisions GPUs across RunPod, Vast.ai, PRIME Intellect, and Lambda Labs, then runs your ML experiment scripts on them with one command. Point it at a shell script like Karpathy's autoresearch example, pick H100s or whatever, and it handles querying real-time pricing, spinning up instances in parallel, distributing runs round-robin, streaming output live, and aggregating metrics into a report. Works across providers without manual SSH or API fiddling—install via pip or download cli tool from the GitHub repo.

Why is it gaining traction?

It stands out by automating the messy multi-cloud GPU hunt: interactive tables for cheapest offers filtered by bandwidth/region, auto-mode for headless runs, resume from interruptions, and network volumes to skip reinstalls. CLI commands like `autofoundry inventory -g H100`, `run script.sh --auto`, or `status` make it dead simple versus scripting each provider's API. Pairs nicely with cli github actions or cli github copilot for workflows on cli github linux or cli github ubuntu.

Who should use this?

AI researchers grinding hyperparameter sweeps or model benchmarks across GPU types/providers. Teams testing autoresearch-style nanoGPT training on spot H100s without per-cloud boilerplate. Python devs on cli tools ai who want cli tools for mac, cli tools for windows, or cli tools linux to offload experiments cheaply.

Verdict

Grab it if you're tired of manual GPU provisioning—solid for quick experiments despite 13 stars and 1.0% credibility score signaling early days. Docs are crisp with demos, but expect occasional provider quirks; test small before big runs.

(198 words)

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.