Skip to content

feat(ST-38): generate safe original memory songs - #60

Merged
afaryy merged 3 commits into
mainfrom
feat/ST-38
Sep 3, 2026
Merged

feat(ST-38): generate safe original memory songs#60
afaryy merged 3 commits into
mainfrom
feat/ST-38

Conversation

@afaryy

@afaryy afaryy commented Sep 3, 2026

Copy link
Copy Markdown
Owner

Summary

  • Add a consent-scoped original-memory-song path using Lyria 3 with Vertex AI service-account authentication in Cloud Run.
  • Reject requests for artist, song, or voice imitation; offer instrumental and no-music fallbacks.
  • Mix temporary generated audio into fixed 60-second photo/video films, looping as needed and fading out safely.
  • Expose simple soundtrack choices in the web flow and document the hosted-verification boundary.

Verification

  • API suite: 59 passed. One existing Starlette/httpx deprecation warning remains.
  • Web suite: 16 passed.
  • Production web build: passed.

Hosted verification still required

Lyria 3 is a preview service. This PR is locally implemented and tested, but it is not production or competition evidence until a non-sensitive hosted smoke test produces a playable generated-audio render.

@afaryy
afaryy merged commit b7d0a8d into main Sep 3, 2026
6 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant