Most transcripts are mostly right. SubtitleGenerator turns the uncertain words into a short review queue—so you fix what matters, then style and export.
THE FIX LOOP
Review only what matters
Other tools dump a full transcript and leave you hunting. After AI captions your video, we flag the words we are unsure about and put them in a focused queue. Clear the list—then you are done.
See the flags — uncertain words are highlighted in the subtitle list and on the timeline.
Jump the queue — move flag by flag (Tab and the review controls) instead of scrubbing the whole video.
Hear the context — play the cue, type a correction, accept a suggestion, or keep the original wording.
Reach All clear — when the queue is empty, you have reviewed the risky bits without rereading every line.
Fix Loop is available on Free and Pro—it is the core workflow, not a paid add-on. You can still edit any cue, split or merge lines, and trim timings whenever you need.
Change caption density (Short / Standard / Clean) to rebuild line breaks from the stored transcript on your device—no extra processing minutes.
Undo / Redo restores caption edits locally. It does not refund minutes already used for a finished AI request.
1. Start a project
Open the editor from the homepage (no signup required to try). Drop or choose a video: MP4, MOV, WebM, or MKV. Your original file stays in the browser for preview and export; only an audio extract is sent for speech-to-text when you run transcription.
Projects can autosave as a draft on this device (local storage). Use Continue editing on the homepage to resume, or start a new video when you are ready.
2. Spoken language & High Accuracy
Before you transcribe, set the spoken language of the audio (default English). Pick the language people are actually speaking—there is no Auto mode. If the first pass looks wrong, change the language and run transcription again.
High Accuracy (Pro / Max) uses a stronger speech model when the take has to be right. It uses 2× processing minutes for that audio length. Turn it on before you transcribe or before a range re-transcribe.
3. Style, animate, and canvas
Every plan can preview all built-in caption styles, animation presets, fonts available on Free, layout, and canvas ratios (for example 9:16, 16:9). Position captions and use the safe zone so text stays readable on Shorts and YouTube.
Saving custom brand styles and uploading custom fonts are Pro / Max features. Styling and preview do not consume processing minutes.
4. Pro tools: Range, Translate
On Pro or Max you can also:
Range re-transcribe — select a messy time range and re-run speech-to-text on that slice only (uses processing minutes for the range length; High Accuracy still doubles that cost).
Translate — send the subtitle text (not the video) to translation and replace or add a translated track. Translation uses processing minutes once per run.
Buying a pay-as-you-go minute pack adds capacity only. It does not unlock these Pro capabilities—you still need a Pro or Max subscription for HD export without watermark, subtitle files, Range, Translate, and High Accuracy.
5. Export
Export is produced in your browser. It does not use processing minutes.
Free / guest — burned-in video around 720p with a small SubtitleGenerator watermark.
Pro / Max — HD (about 1080p) without watermark, and downloadable subtitle files (SRT, VTT, and other common formats).
Depending on your browser, burned-in video may be offered as WebM and/or MP4. Prefer the format your publish platform accepts.
6. Processing minutes & plans
Processing minutes pay for AI work: transcription, High Accuracy, range re-transcribe, and translation. Guests can try a few transcriptions without an account. Sign in for a Free monthly minute allotment, or upgrade for more minutes and Pro features.
Pro / Max — more monthly minutes, longer videos, HD / no watermark, subtitle files, Range, Translate, High Accuracy. Max is the same toolkit with higher capacity.
PAYG packs — one-time minutes that do not expire; subscription minutes are used first when both are available.