Transcribe a screen recording on your Mac and edit it by the word
Qreata turns the narration of a screen recording into a transcript without sending the audio anywhere. The transcript is an editing tool: cut the recording at word level, remove the silences and the filler words, and burn word-highlighted captions into the video.
7-day trial · $49 once · no subscription · Apple Silicon · macOS 14+The short answer
Qreata transcribes the speech in a screen recording locally on your Mac with a bundled Whisper model; the audio is not sent anywhere and no network is needed. In the Script panel you cut the recording at word level, click Remove silences and Remove filler words (English and Russian speech), and turn on Captions in video to burn word-highlighted captions into the MP4. It does not export SRT or VTT caption files, and the caption style is fixed.
Use it when
The narration needs tightening and the recording should stay on your Mac.
- Your narration has false starts and repeated sentences, and you would rather find them in text than scrub through audio.
- The recording shows things you won’t upload to a transcription service.
- You want an 8-minute tutorial to lose its dead air without a second take.
- Viewers watch muted, in a help center or a chat preview, and captions in the picture are enough.
It won’t help when
You need a caption file or a different sentence from the one you said.
- You need SRT or VTT files, or captions in your own font and colors. Captions are burned in with one fixed style.
- You said the wrong thing and need the right words. A cut can’t add them; say the sentence again with Record voiceover · keep video, or use Replace Take.
- The audio needs noise removal. Qreata has no noise reduction.
- You record on Windows or an Intel Mac.
How local transcription works in Qreata
Transcription runs on your Mac in the editor’s Script panel. The Whisper model ships inside the app, works without a network connection and detects the spoken language automatically. The transcript ties each spoken word to its place in the recording.
That link is what you edit with. A word-level cut removes one word or phrase, video and audio together, instead of a rough slice of the timeline. Remove silences takes out pauses of 0.8 seconds or longer, and Remove filler words takes out “um” and “uh” in English and Russian speech. Captions in video draws the same words onto the picture, highlighted as they are spoken.
Example: tightening an 8-minute tutorial from its transcript
You record an 8-minute tutorial in one take. It is accurate, but you pause for a long time while pages load, you say one sentence twice at 2:10, and at 5:20 you start a sentence wrong: “Click Export, sorry, click Share.” The cleanup looks like this:
- Open the Script panel and let the recording transcribe on your Mac.
- Click Remove silences to take out the dead air, then Remove filler words: “um” and “uh” go automatically, and words such as “like” or “actually” are removed only if you choose to.
- Find the repeated sentence at 2:10 in the transcript and cut the first attempt.
- Go to 5:20 in the transcript and cut “Click Export, sorry,” at word level. The sentence now starts at “click Share.”
- Play each cut with the screen in view. A cut removes that moment from the video as well as the audio, so watch for a cursor that jumps.
- Undo any cut that clips a consonant or makes the sentence sound rushed. Edits are non-destructive.
- Turn on Captions in video if the tutorial will be watched muted, and play a minute to check product names in the caption text.
- Export the MP4.
What transcription does and doesn’t include in the beta
The feature covers review, cutting and captions drawn into the picture. Caption files are not part of it.
| Task | In the beta | Notes |
|---|---|---|
| Transcribe speech on the Mac | Yes | Bundled Whisper model, no network, language detected automatically |
| Cut at word level | Yes | Removes video and audio together; undo works |
| Remove silences and filler words | Yes | Pauses from 0.8 s; “um” and “uh” in English and Russian, “like” or “you know” opt-in |
| Burn captions into the video | Yes | Captions in video: word-highlighted, fixed style |
| Copy the transcript text | Yes | Paste it into a help article or a ticket reply |
| Export SRT or VTT caption files | No | Use burned-in captions, or caption the MP4 in another tool |
| Remove background noise | No | Qreata has no noise reduction |
Word-level cut or Replace Take?
A cut removes words. It can’t change them. Choose by what the sentence needs:
- A false start, an aside or a repeated phrase: word-level cut.
- Long pauses and “um”s between sentences: Remove silences and Remove filler words.
- A sentence that names the wrong button while the screen shows the right one: Record voiceover · keep video, so you say it correctly over the same picture.
- A wrong click with correct narration: Rerecord video · keep audio for that range, or Replace Take to redo picture and sound together.
- Thirty seconds of rambling between two good sections: split and delete in the editor.
Captions: burned into the video, no SRT or VTT file
Captions in video burns captions into the picture, with each word highlighted as it is spoken. They are part of the exported MP4, so they show in every player, including a helpdesk widget or a chat preview that has no caption support. The style is fixed: you don’t choose the font, the colors or the position.
Qreata does not export SRT or VTT. If your platform or an accessibility policy requires a caption file the viewer can switch on and off, export the MP4 and caption it in a separate tool. That is the one case where this feature is not enough. You can copy the transcript text out of the Script panel as a starting point.
Transcripts also misread unusual product names. Play the audio around a cut before you trust the text. A 10-minute check in the 7-day trial: record 2 minutes with your product names and a deliberate false start, transcribe, cut it, turn on Captions in video, and watch the result.
Questions worth asking
Is the transcription done on my Mac?
Yes. Qreata transcribes locally with a Whisper model that ships in the app. It works without a network connection, and the audio is not sent to a transcription service.
Can Qreata add captions to a screen recording?
Yes. Captions in video burns word-highlighted captions into the exported MP4. The style is fixed, and the captions can’t be switched off by the viewer.
Can I export SRT or VTT subtitles?
No. Qreata burns captions into the video but does not export SRT or VTT files. If you need a caption file, caption the exported MP4 in a separate tool.
Can I remove pauses from a screen recording automatically?
Yes. Remove silences in the Script panel cuts pauses of 0.8 seconds or longer, and Remove filler words takes out “um” and “uh” in English and Russian speech. Replay the result, because a pause after a click gives the viewer time to see the screen change.
Can I edit the video by editing the transcript?
Yes, for cuts. You can remove words and phrases at word level. You can’t change what you said; for that, say it again with Record voiceover · keep video, or re-record the range with Replace Take.
Is transcription included in the price?
Yes. It is part of the app: the Personal license is $49 once for 3 Macs, and the 7-day trial includes every feature.