SubtitleGenerator combines transcription, uncertain-word review, timing edits, styling and export in one browser editor. I am the maker, so this is a product post, not an independent review.
The correction step is the core of the workflow: uncertain words are flagged, and each cue can be kept, edited or re-transcribed. A visible remaining count gives the flagged queue a stopping point, but a final listen-through is still necessary. Confidence flags cannot catch every error.
The original video stays on the device; only the extracted audio is sent for transcription. That reduces the need to upload the source video, but transcription is not fully offline.
Free to start, with paid options. Built for creators and editors who want to finish corrections before styling and exporting.
Feedback I am looking for: does the correction workflow make it clear what has been reviewed and what still needs a human check? No votes or endorsements requested.
Disclosure: maker post; AI-assisted drafting.