Skip to content
Get packs

Vast.ai GPU Rental

Rent, set up, and destroy cloud GPU instances on vast.ai for your training jobs

Source Download skill
Repository
wanshuiyin/Auto-claude-code-research-in-sleep
Stars
18,482
Licence
MIT
Files
1 file

What it does

This skill lets an AI agent rent on-demand graphics cards (GPUs) from the vast.ai marketplace for you. Instead of picking hardware yourself, you describe the training task; the agent estimates the memory and time needed, searches for offers, shows you a few options with estimated total cost, and after you choose, creates the rental, sets it up, and later destroys it. It also keeps a local JSON file of active instances so other experiment skills can connect to them.

Great for

  • Researchers or engineers who need a GPU for a training run but do not own one
  • People running multi-experiment workflows who want cost estimates before renting
  • Users already comfortable with vast.ai accounts, API keys and SSH keys

Skip it if

  • Anyone who does not want to spend real money: renting instances is billed to your vast.ai account, and forgotten instances keep charging until destroyed
  • Sensitive or regulated data: the rented machines belong to third-party hosts, so data placed on them leaves your own environment
  • Users unfamiliar with the command line, since setup requires installing a Python CLI and handling API keys

Try asking

  • “Find me a cheap GPU to fine-tune a 7B model on my dataset for about 3 hours”
  • “Rent a server on vast.ai for the baseline experiment in my plan”
  • “Show me GPU options for training ResNet-50 and estimate the total cost”
  • “Destroy the vast.ai instance for experiment exp01 now that it has finished”
Step-by-step guide: how to use Vast.ai GPU RentalInstall, first run, examples, tips and FAQs

Install in 3 steps

  1. 1

    Download the skill

    The ZIP contains the 1 original files, the MIT licence and an attribution note.

    vast-gpu.zip
  2. 2

    Add it to your agent

    Claude Code (all projects)

    mkdir -p ~/.claude/skills && unzip vast-gpu.zip -d ~/.claude/skills/

    Claude Code / Agent SDK (this project only)

    mkdir -p .claude/skills && unzip vast-gpu.zip -d .claude/skills/

    Claude apps (web/desktop): Settings → Capabilities → Skills → Upload skill, and choose the ZIP.

  3. 3

    Ask for the task

    No special command needed — the agent reads the skill's description and loads it when your request matches. Try one of the prompts above.

Get better results

  • Upload your SSH key to vast.ai before renting, because keys are only applied at creation and adding one later means destroying and recreating the instance.
  • Always check the estimated total cost table and destroy instances as soon as work finishes to avoid unexpected charges.
  • Give the agent a precise task description (model size, dataset size, duration) so the VRAM and hours estimates are reasonable.

What's inside

  • SKILL.md

Redistributed unmodified from wanshuiyin/Auto-claude-code-research-in-sleep under the MIT licence. The assessment was drafted with AI from the skill's files on 2026-10-29. Skills can instruct your agent to run code — read SKILL.md and any scripts before installing.