Community-contributed example Factbooks for the Factlet Protocol. Real artifacts that implementers and users can read, fork, and learn from.
| Domain | Path | Highlights |
|---|---|---|
| Payments (Stripe) | examples/payments |
money handling, refund workflow, supersession example |
| React + TypeScript frontend | examples/frontend |
retired-tech (Redux), a11y floor, bundle budget |
| ML training pipeline | examples/ml-pipeline |
data lineage, eval gates, PII handling |
Each directory contains <domain>-factbook.yaml + README.md explaining what the example illustrates and how to use it. Filenames follow the <scope>-factbook.yaml convention so the project scope is visible without context.
See CONTRIBUTING.md. What's expected of a registry contribution:
- Schema-valid against the v0.1 spec
- No secrets, no PII, no proprietary content
- Sanitized example — fictional but realistic
- README explaining the domain, what's interesting about the facts, and how to test the AI's compliance with them
- 5-10 factlets minimum, focused on facts an AI would otherwise get wrong
Once the reference SDK ships, run:
factlet validate examples/<your-domain>/<your-domain>-factbook.yamlFor now, validate manually against the schema.
Examples in this repository are MIT-licensed. See LICENSE.