This repository publishes downloadable voice model packs used by Happier native apps.
Packs are published as:
- a
manifest.jsonper pack (consumed by the app) - a set of files referenced by that manifest (downloaded + sha256 verified on-device)
In the app, configure:
EXPO_PUBLIC_HAPPIER_MODEL_PACK_MANIFESTSas a JSON map{ [packId]: manifestUrl }
npm run build:model-packs -- --repo happier/happier-assets --tag model-packs-devArtifacts are written to:
dist/packs/<packId>/...(human-inspectable pack tree)dist/release/<asset files>(release-ready asset names)
Use the GitHub Actions workflow Release model packs (manual dispatch) or push a tag matching model-packs-*.