Installation

The core framework needs Python 3.11+ and Git. It does not require a GPU. External tasks, model endpoints, learner CLIs, and sandbox support depend on the workflow you choose.

Install the checkout

Recommended for examples, benchmark recipes, and development.

If you do not have a checkout:

git clone https://github.com/Infini-AI-Lab/hermit.git
cd hermit

From the repository root:

python3 -m venv .venv
source .venv/bin/activate
python -m pip install -e .
hermit --help

Help should list check, train, eval, deploy, convert, and serve. Activate the environment again in each new terminal. Editable installation uses this checkout’s Python code.

Install from PyPI

Install the core framework in a virtual environment:

python3 -m venv .venv
source .venv/bin/activate
python -m pip install hermit-harness
hermit --help

The package includes the CLI, adapter templates, and built-in Evolvers. GEPA, SkillOpt, and Harness² also need their method-specific dependencies. The example data, benchmark recipes, and requirements-*.txt files live in the repository; use a checkout for those workflows.

Choose the additional requirements

Workflow	Additional requirements
Offline first training	None beyond the core install
First real-model training	python -m pip install -e . -r requirements-gepa.txt, a compatible chat-completions endpoint, model id, and credential; data is included
Generate a Codex / Claude Code / Pi workspace	No host CLI is needed until you start the host
Run an acting host	Installed and authenticated host, plus adapter tool dependencies
Prompt-only learner	The Codex CLI, Linux with bubblewrap, and a learner backend that serves the Responses API
GEPA	python -m pip install -e . -r requirements-gepa.txt, learner endpoint, validation split
SkillOpt	python -m pip install -e . -r requirements-skillopt.txt, learner endpoint, validation split
Codex / Claude learner	Selected authenticated CLI and supported program-execution environment
MWM	Linux learner environment, Codex, evidence store, task deployment contract, model access
Benchmark recipe	Its dataset, scorer, model settings, and environment; see catalog

Core installation and the deterministic tutorial work on macOS. Sandboxed learner programs use a Linux-specific bubblewrap backend. For Ubuntu/Debian:

sudo apt-get install bubblewrap
bwrap --version

The environment must permit the required namespaces. Installing the binary alone may not satisfy a container or cluster’s namespace policy. The sandbox does not isolate your acting pipeline or arbitrary in-process Python. See sandbox boundaries.

A recipe can require Python 3.12+ even though the core supports 3.11. Check external dependencies separately. For exact experiment reproduction, record the framework commit and dependency revisions.

Verify the local installation

From the source checkout:

hermit check --config docs/_examples/first_training/run.yaml

The probe should score the first toy task zero without execution errors. Its sandbox-disabled warning is expected for the included in-process, offline teaching learner. Continue to Quick Start to train and evaluate it.

Optional: build documentation

python -m pip install -r docs/requirements.txt
bash docs/build.sh
bash docs/serve.sh 8000

Open http://127.0.0.1:8000. The preview rebuilds on source changes. The strict build treats warnings as errors. The local preview includes design drafts under _static/drafts/; the default publication build excludes them. For earlier evolve / run interfaces and current limitations, see Compatibility.