FAQ¶

What is START?¶

A pipeline for research-driven, personalized curricula using Perplexity and OpenRouter.

How do I install and run tests?¶

See Environment: ./environment.md. Then run:

uv sync --all-extras --dev
uv run pytest -q

How do I generate content?¶

See Getting Started: ./getting_started.md and Pipeline: ./pipeline.md.

Where do outputs go?¶

Research in data/audience_research/ and data/domain_research/; curricula in data/written_curriculums/; visuals in data/visualizations/; translations in data/translated_curriculums/. See Data & Outputs for directory purposes and artifact descriptions.

How do I see the docs site?¶

Run ./run_docs.sh --serve (local) or visit GitHub Pages at the project site.

How can I contribute?¶

See Contributing: ./CONTRIBUTING.md.

Where are examples?¶

See Examples: ./examples.md and tests in tests/.

What keys do I need?¶

Required: PERPLEXITY_API_KEY and OPENROUTER_API_KEY. Optional: OPENAI_API_KEY. Copy .env.example to .env and fill in values.

Do tests use the network?¶

No external provider calls run by default. Provider behavior is exercised with local HTTP servers; live connectivity requires credentials and an explicit probe invocation. See Testing.

Docs site won’t start — what should I check?¶

Use ./run_docs.sh --serve. Ensure mkdocs.yml exists at repo root. The script runs MkDocs through the locked project environment with uv.

Visualizations not generating?¶

Run uv run start-curriculum --non-interactive --stages visualizations --json. If running headless (servers/CI), set MPLBACKEND=Agg.

CLI examples don’t work from my cwd¶

Use the module entrypoints from the repository root; numbered scripts remain available for staged workflows.