OpenParable
Skip to content
← Back to skills

Modal Compute

Apps & Automation
modal-compute

Modal Compute is an Agent Skill for apps & automation published in the companion-inc/feynman repository on GitHub: a SKILL.md instruction file an AI agent can load to do this task.

Run explicitly chosen research benchmark or replication jobs on Modal's serverless infrastructure. Use when a Feynman research workflow needs burst remote GPU compute and the Modal CLI is available.

View on GitHub ↗
QUICK START

How to use this skill

Bring this guide into your coding agent with a prompt tailored to the tool you use.

  1. Open your project in Codex.
  2. Copy and paste the prompt below.
  3. Review files and risks before approving any action.
Prompt to paste
Inspect the Agent Skill at https://github.com/companion-inc/feynman/blob/HEAD/skills/modal-compute/SKILL.md and all supporting files. Report provenance, permissions, scripts, dependencies, and security risks. Propose an exact installation into .agents/skills/modal-compute, but wait for my approval before writing files or running scripts.

Copying this prompt does not install or run the skill. Review third-party files before use.

Modal Compute

Use the modal CLI for bounded research experiments that need burst GPU compute. No pod lifecycle to manage; write a decorated Python script, run it, and save raw outputs back into the research artifact folder. Do not use this skill to deploy services or unrelated batch jobs.

Setup

pip install modal
modal setup

Commands

CommandDescription
modal run script.pyRun one research experiment script on Modal
modal run --detach script.pyRun a long research experiment and record the returned app/run identifier
modal shell --gpu a100Open an interactive GPU shell for research environment debugging

GPU types

T4, L4, A10G, L40S, A100, A100-80GB, H100, H200, B200

Multi-GPU: "H100:4" for 4x H100s.

Script pattern

import modal

app = modal.App("experiment")
image = modal.Image.debian_slim(python_version="3.11").pip_install("torch==2.8.0")

@app.function(gpu="A100", image=image, timeout=600)
def train():
    import torch
    # training code here

@app.local_entrypoint()
def main():
    train.remote()

When to use

  • Bounded replication or benchmark jobs that need burst GPU
  • No persistent state needed between runs
  • Check availability: command -v modal