Getting Started¶

This guide is the shortest verified path from a fresh clone to a useful RoboCI run. After RoboCI is installed from a package, roboci demo, roboci demo --openloop, and roboci benchmarks ... also work outside a repository checkout because their small demo assets are bundled with the package.

Requirements¶

• Python 3.11 or newer.
• Docker Compose for the default dashboard path.
• Node.js 20 or newer only if you want to develop the dashboard locally without Docker.
• A shell that can create a virtual environment and run local commands.

Install¶

git clone https://github.com/GagandeepReehal/roboci.git
cd roboci
python3.11 -m venv .venv
source .venv/bin/activate
python -m pip install --upgrade pip
pip install -e ".[dev]"

Verify the package and CLI:

python -c "import roboci; print(roboci.__version__)"
python -m roboci --help
roboci --help

Run The Demo¶

roboci demo
roboci report --run runs/latest

Expected output: RoboCI prints paths for a passing smoke baseline, an expected collision failure, a current smoke run, regression output, and generated report.html and report.md files.

Run the smoke benchmark directly:

roboci benchmarks run smoke

The demo and benchmark commands use packaged assets. Commands that pass explicit paths such as examples/suites/smoke.yaml still expect those files to exist in the current checkout.

Run The Dashboard¶

docker compose --profile dashboard up --build api dashboard

Open http://127.0.0.1:3000/dashboard. The API is published at http://127.0.0.1:8000.

For local dashboard development without Docker:

cd frontend
npm install
cd ..
roboci dev

If you do not have Node locally but want the same entry point, run roboci dev --docker.

Next Steps¶

• See Examples for closed-loop, open-loop, failure, and regression workflows.
• See Configuration for environment variables, ports, paths, and suite configuration.
• See CLI for every command and exit-code behavior.