[vedanth-jadhav]

Problem

Caption translation starts the server action inside a React state updater. Under Strict Mode two simultaneous requests for the same language can trigger four translation calls; concurrent languages also clear the loading indicator as soon as the first request finishes. A rejected server action leaves its caller's promise unresolved, and a late result for one video can populate the next video's caption cache.

Changes

• Start translation outside the state updater, reuse in-flight requests per language, and keep the loading indicator active until all requests settle.
• Resolve failed requests to null so callers can retry, and cache successful translations synchronously.
• Scope the provider to the video ID and ignore state updates after unmount. Changing videos remounts the provider's children along with their caption state.

Validation

• 19 focused tests pass, including five mounted Strict Mode regression cases and existing transcript tests. The duplicate request, concurrent loading, and video-switch cases fail against the current main branch.
• Scoped strict TypeScript check of the provider and tests, using declarations matching the two server-action signatures; Biome and workflow YAML checks pass.

The tests mock the translation server actions; they do not exercise a live AI provider or a full web build.

Confidence Score: 5/5

The PR appears safe to merge; no actionable issues were found.

Summary

This PR moves caption translation requests out of React state updaters and keeps each video's caption state separate.

• Caption requests share translations and stay separate for each video.

No actionable issues were found. Tests were inspected, not run during this review.

_{Reviews (1) · Last reviewed commit: "Fix caption translation request ownershi..." · Reviewed by Greptile}

[vercel]

Someone is attempting to deploy a commit to the Cap Software Inc Team on Vercel.

A member of the Team first needs to authorize it.