Skip to content

fix: refuse text-only local video models for footage production at start (#10457) - #10477

Merged
atomantic merged 1 commit into
mainfrom
claim/issue-10457
Oct 7, 2026
Merged

atomantic merged 1 commit into
mainfrom
claim/issue-10457

Conversation

@atomantic

Copy link
Copy Markdown
Owner

Footage production always animates a frame, so a local video model without image-to-video support can never produce a clip, yet Start accepted it and only failed at the first clip.

  • routeEligibility now refuses a local video model lacking image in its supported modes, so POST /:id/production-runs returns 409 naming the model and the missing capability.
  • POST /api/music-video/autonomous refuses a pinned text-only video:local model with a 400 (VIDEO_MODEL_TEXT_ONLY) before creating the project.
  • The start drawer's local-video picker no longer lists text-only models.

Remaining

Letting a parked run swap its video model and continue without cancelling is split to #10473.

Closes #10457 is intentionally not used.

Refs #10457

@atomantic
atomantic merged commit 5302ee2 into main Oct 7, 2026
10 checks passed
@atomantic
atomantic deleted the claim/issue-10457 branch October 7, 2026 05:48
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.

Music Video production accepts a text-only video model and only fails at the first clip (no image-to-video)

1 participant