Skip to content

feat: reprocess saved transcripts with AI cleanup#1143

Draft
adamholter wants to merge 1 commit into
OpenWhispr:mainfrom
adamholter:codex/reprocess-transcript
Draft

feat: reprocess saved transcripts with AI cleanup#1143
adamholter wants to merge 1 commit into
OpenWhispr:mainfrom
adamholter:codex/reprocess-transcript

Conversation

@adamholter

Copy link
Copy Markdown
Contributor

What changed

  • adds a History action that reruns AI cleanup against the exact saved raw transcript
  • updates the existing row only after valid reasoning output and a successful database write
  • preserves raw text byte-for-byte and keeps failures non-destructive
  • surfaces unavailable-provider and persistence errors without optimistic UI drift
  • adds all current locale strings and focused provider/database/UI coverage

Why

Retry currently covers transcription failures, but users cannot deliberately rerun cleanup after changing models, prompts, or settings. This adds that narrow workflow without retranscribing audio or creating duplicate history rows.

Validation

  • 263 passing tests, 0 failures, 5 macOS-expected Linux skips
  • focused provider failure, empty output, missing raw text, and database failure coverage
  • formatting, TypeScript, i18n, and renderer build pass
  • headless fail-then-success flow passed with exact write-payload assertions and zero console errors

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