Set up and transcribe
- Open a video or audio file, then Editing workspace > Audio > Speech to subtitles. Choose a results folder.
- Choose Fast or More accurate recognition quality and download its model once. Set Arabic, English or Detect automatically.
- Choose Speech for spoken audio or Singing & music for songs and processed vocals. Start Transcribe speech and follow the task.
The full open file is transcribed in its spoken language. This is separate from AI translation. Model downloads require internet. The recording is processed on your computer. Singing is harder to recognize, so review words and timing.
Add the result and correct it
- On the completed task, choose Use transcript. Captions are added to the project and timeline. If no speech was recognized, try the other content mode, a chosen language or a more accurate model.
- Open Captions > Edit text & timing. Click a cue time to seek to it, then correct the text.
- Expand Find, replace & adjust timing to preview matches, replace selected or all matches, or shift timing. A negative shift moves captions earlier. Select cues to merge, or split a cue at the playhead.
Empty or invalid results do not replace existing captions. Check overlapping or invalid timings before export.
Export subtitles and review individual words
- Open Export subtitles & word timing. Save an SRT file for common players or VTT for web use.
- If recognition produced word timings, expand Edit word timing and correct the text, start and end of each word.
- In Word-by-word captions, enable Show synchronized words and choose Outline, Readable box or Pop in. Review playback before exporting.
Captions in an audio project
Use Preview in editor to listen while reading captions. Create video with subtitles opens video output for the audio and its visualizer. Captions follow kept ranges, cuts and speed changes, and can be included or hidden during video export. MP3 and WAV contain sound only. Save SRT/VTT separately or export a video to keep visible text.