Atomsh

Your autonomous coding agent for science.

Atomsh reads and edits files, searches a codebase, and runs commands from your terminal. It also queries JARVIS-DFT, runs ALIGNN predictions, computes band structures and analyses diffraction patterns, in the same loop, so it can look a material up instead of recalling it.

It runs on AtomGPT alone. One account, no provider keys to configure.

You need an atomgpt.org account

That account is the only credential Atomsh uses. Sign up at atomgpt.org first, then install.

Install

curl -fsSL https://atomgpt.org/install | bash

Then connect and start:

atomsh login
atomsh

What it looks like

atomsh v2026.9.2, your autonomous coding agent for science
gemma-4-26b · /home/you/project · materials + code · /help for commands

› calc.py crashes when you run it. Fix it and verify.
  · read_file(path=calc.py)
  · edit_file(path=calc.py, old_string=def divide(a, b): …)
  · bash(command=python3 calc.py)
Fixed the ZeroDivisionError in calc.py:6 and verified by running it.

And the part no other coding agent does:

› bandgap of silicon JVASP-1002 from JARVIS-DFT
  · explore(app=/jarvis_dft/query, params={"jid": "JVASP-1002"})
OptB88vdW 0.731 eV, mBJ 1.277 eV, HSE 1.22 eV

Why it exists

A general coding agent asked about ALIGNN will describe something plausible and wrong, because it is answering from model weights. Atomsh carries the AtomGPT tools, so it answers from the database instead.

Where to go next

• Getting started for install and sign-in details
• Using Atomsh for the commands and flags
• Tools for what it can actually do
• Permissions for what it will ask before doing

Requires Python 3.10 or newer. Linux and macOS; Windows through WSL. Apache-2.0 licensed.

Citing

Atomsh builds on the AtomGPT platform and the JARVIS infrastructure. If it contributes to work you publish, please cite the relevant papers.

1. J. Lee, J. Ely, K. Zhang, A. Ajith, C. R. Campbell and K. Choudhary, "AGAPI-Agents: An Open-Access Agentic AI Platform for Accelerated Materials Design on AtomGPT.org", The Journal of Physical Chemistry Letters 17 (26), 7221-7231 (2026). doi:10.1021/acs.jpclett.6c00837

2. K. Choudhary, "The JARVIS infrastructure is all you need for materials design", Computational Materials Science 259, 114063 (2025). doi:10.1016/j.commatsci.2025.114063

3. K. Choudhary, "ChatGPT Material Explorer: Design and Implementation of a Custom GPT Assistant for Materials Science Applications", Integrating Materials and Manufacturing Innovation 14 (3), 276-283 (2025). doi:10.1007/s40192-025-00410-9