SRT caption converter, both directions

SRT is the lingua franca of captions - YouTube, Instagram, LinkedIn, and every serious editor accept it. But working with it means two recurring chores: stripping timestamps out of an SRT to get readable text, and adding timestamps to a script so a platform will accept it as captions.

This converter does both in one box. Paste either format and it detects the direction automatically: SRT in, clean transcript out; transcript in, properly numbered and timed SRT out.

Turning text into captions is mostly about pacing. The converter packs your transcript into cues of at most ~84 characters, splits each cue into balanced lines that fit a vertical video, and never merges text across paragraph breaks - a new paragraph starts a new cue.

Durations come from word count at your chosen reading pace (slow, natural, or fast narration), clamped between 1.2 and 6 seconds. The result uploads cleanly and reads comfortably, even if nobody ever sees the timestamps.

Caption files are often the full script of unreleased content, so this tool does the conversion entirely in your browser: no upload, no server-side processing, no account. Close the tab and nothing persists anywhere.

A words-per-minute estimate is fine for a talking-head upload, but word-by-word caption sync - the karaoke highlight that holds attention on short-form video - needs real timestamps from real audio. That is what FrameCompose generates: an AI voiceover transcribed with per-word timing, captions placed automatically, and every caption editable on the timeline afterwards.

Paste your transcript, click "Make a captioned video from this", and the studio opens with your script pre-loaded.