Mesocosm CLI

Central hub for the mesocosm command-line tool — SWECC’s BenchAnything / Mesocosm platform.

New here?

Event logistics → SWECCATHON 2026 slides. Then the platform tour → mesocosm overview slides.

Install with pip (recommended — the package is published on PyPI):

pip install swecc-mesocosm

Use pip, not Homebrew, uv, Poetry, or other wrappers, unless you know what you are doing. Those can install a different Python environment or an outdated build and cause confusing mesocosm: command not found errors. No backend source code required.

Quick start

python3 -m venv .venv
source .venv/bin/activate   # macOS/Linux — Windows: .venv\Scripts\activate
 
pip install swecc-mesocosm
mesocosm --help
 
mkdir my-env && cd my-env
mesocosm init
 
ollama serve                # or open the Ollama app
ollama pull llama3.2
python adapter.py           # terminal 1
mesocosm run local          # terminal 2

Platform:

mesocosm auth login
mesocosm env submit --name "My env" --github-url https://github.com/you/your-repo
mesocosm run create --domain YOUR_DOMAIN_ID --vow-version 1.0.0 --model gemini/gemini-3.1-flash-lite

Guides

Who this is for

  • Environment authorsbenchanything.json, env.py, adapter
  • Platform users — runs, teams, showcase exports

Production vs local

ModeTypical use
Production (default)mesocosm auth login, env submit, run createhttps://api.swecc.org
Localexport MESOCOSM_LOCAL=1, mesocosm doctor --local, mesocosm run local with Ollama

Details: Configure · Troubleshooting

CLI help

mesocosm --help
mesocosm auth --help
mesocosm run --help

Every flag and default: Command reference.