APPRENTICE / QUICKSTART

Your first
extra pair of hands.

Get Apprentice from the GitHub repository, then try your first workflow with the local mock finance portal.

Prefer a download? Get the Phase 5 source ZIP, including tests, examples, and development documentation.

1. Set up on Windows.

Install Python 3.11 or newer and Git. Open PowerShell and clone the repository:

git clone https://github.com/Villain45/Apprentice.git
cd Apprentice

Alternatively, extract the source ZIP and open PowerShell in its Apprentice folder. Then run:

python -m venv .venv
.\.venv\Scripts\python.exe -m pip install -r requirements.txt
.\.venv\Scripts\python.exe -m playwright install chromium

2. Open the mock portal.

Keep this terminal open:

.\.venv\Scripts\python.exe -m http.server 8000 --directory src/apprentice/portal

3. Preview the workflow.

In a second PowerShell terminal, preview the included reviewed task. Preview creates no browser or run files.

.\.venv\Scripts\python.exe src/apprentice/executor/run_workflow.py workflows/monthly_report.yaml

4. Let it take a turn.

Run with fresh demo inputs. The tool asks privately for a password and confirms the workflow, upload, and submission. Use any synthetic password in the mock portal.

.\.venv\Scripts\python.exe src/apprentice/executor/run_workflow.py workflows/monthly_report.yaml --execute --param username=demo --param month=September --param total_amount=1500

5. Teach your own routine.

.\.venv\Scripts\python.exe src/apprentice/recorder/record.py --url "http://localhost:8000/?v=baseline" --name my_task
.\.venv\Scripts\python.exe src/apprentice/labeler/build_workflow.py recordings/my_task/steps.json --output workflows/my_task.yaml

Review the YAML, resolve every uncertain step, and validate it before replay. The included README and DEVELOPMENT files explain browser teaching, system capture, terminal enrollment, and reviewed file operations.

Prefer a project task?

.\.venv\Scripts\python.exe src/apprentice/executor/run_workflow.py workflows/project_tests.yaml

The sample opens an Explorer directory, verifies an isolated VS Code workspace, and runs that project's tests using an explicit Python interpreter. Read the README before its first execution.

See the demos ↗