One concept: a GitHub Actions workflow runs your existing suite on a fresh Linux box on every push and pull request. One skill: reading the workflow file, putting it in the one folder GitHub scans, and reading the Actions result.

CI is a recipe GitHub follows automatically. When you make your repo from the ichiramen-dojo-py template, the workflow already ships inside it - you do not write it from scratch. The instant you push, GitHub spins up a clean Linux machine, installs Python and the browsers, runs pytest, and shows a green check or a red X in the repo's Actions tab. Nothing new to test; you are just teaching a server to run what you already run. Two things decide whether it runs at all: the file must live in exactly .github/workflows/, and its steps must run in order on a machine that starts empty.

Where the recipe lives. GitHub Actions looks in one exact place: .github/workflows/. Any .yml or .yaml file there is a workflow. The folder name and path matter - GitHub does not scan anywhere else.

ichiramen-dojo-py/
  .github/
    workflows/
      tests.yml        <- GitHub reads this
  requirements.txt
  pytest.ini
  conftest.py
  tests/

The whole workflow, read top to bottom. This is the file that ships in the template at .github/workflows/tests.yml. You do not edit it to get your first run - read it so you understand what CI does on every push.

name: IchiRamen Dojo Tests
on:
  push:
    branches: [main]
  pull_request:
    branches: [main]
jobs:
  test:
    timeout-minutes: 60
    runs-on: ubuntu-latest
    steps:
      - uses: actions/checkout@v7
      - name: Set up Python
        uses: actions/setup-python@v6
        with:
          python-version: '3.12'
      - name: Install dependencies
        run: |
          python -m pip install --upgrade pip
          pip install -r requirements.txt
      - name: Install Playwright browsers
        run: python -m playwright install --with-deps chromium
      - name: Run tests
        run: pytest --tracing=retain-on-failure
      - name: Upload traces
        uses: actions/upload-artifact@v7
        if: ${{ !cancelled() }}
        with:
          name: playwright-traces
          path: test-results/

The five steps, plainly: actions/checkout@v7 copies your repo onto the empty machine; actions/setup-python@v6 installs Python (quote '3.12' so YAML does not read it as a number and drop a trailing zero on something like 3.10); pip install -r requirements.txt installs the same deps you use locally; python -m playwright install --with-deps chromium downloads the browser plus the system libraries it needs to launch on bare CI Linux; pytest --tracing=retain-on-failure runs the suite and keeps a trace for any failure; the upload step ships those traces as an artifact you can download.

Harmless YAML aside. YAML's older spec treats some bare words as booleans - the "Norway problem," where no/yes/on/off parse as False/True. If you ever load this file with Python's yaml.safe_load, the top-level key on: comes back as the boolean True, not the string "on". That is a quirk of that one Python parser; GitHub uses its own workflow parser and reads on: correctly, so your CI is unaffected. Mentioned only so a Python self-check does not throw you.

Four mistakes this rank inoculates you against. The wrong path (ci/tests.yml instead of .github/workflows/tests.yml) is the first - that is beat H2 above, and the file-tree shows the right place. The other three are one-line workflow slips:

      - uses: actions/upload-artifact@v7
        with:
          name: playwright/traces
          path: test-results/

      - uses: actions/upload-artifact@v7
        with:
          name: playwright-traces
          path: test-results/

Artifact names cannot contain / - a slash is rejected and the upload fails.

      - name: Run tests
        run: pytest --tracing=retain-on-failure
      - uses: actions/upload-artifact@v7
        with:
          name: playwright-traces
          path: test-results/

      - name: Run tests
        run: pytest --tracing=retain-on-failure
      - uses: actions/upload-artifact@v7
        if: ${{ !cancelled() }}
        with:
          name: playwright-traces
          path: test-results/

Without if: ${{ !cancelled() }}, a failed test step stops the job and the upload never runs - so the trace of the failure you actually need is never saved. The !cancelled() condition means "run this unless I manually cancelled the whole thing," so traces upload on pass AND on failure.

      - run: python -m playwright install chromium

      - run: python -m playwright install --with-deps chromium

On a bare CI Linux box Chromium cannot launch without its system libraries. --with-deps installs them; leave it off locally (your machine already has them), keep it on in CI.