Third-Party Notices

This repository is MIT licensed. It also depends on third-party Python packages for demos and tests. Their licenses are governed by their respective projects.

Package	Why It Is Used	Notes
anthropic	Anthropic-compatible teaching path and DeepSeek compatibility.	Imported lazily when a real provider is selected.
openai	OpenAI Responses API provider path.	Imported lazily when --provider openai is selected.
python-dotenv	Load local .env files for optional real-provider demos.	No secrets should be committed.
pyyaml	Parse chapter progression metadata and structured teaching files.	Used by tests and verification utilities.
pytest	Offline test suite.	Required for local and CI verification.

The tutorial code does not vendor these packages. Install them with:

pip install -r requirements.txt

Before adding a dependency, prefer the Python standard library for teaching clarity. If a dependency is necessary, document why it exists here and keep provider SDK imports lazy so offline demos remain deterministic.