Sonix's core audience wants the spoken words as editable text — for show notes, a blog repurpose, a quote, a script to translate — not a caption track burned onto the video itself. Versely's transcribe_audio does exactly that distinct job: it runs Cartesia's ink-whisper speech-to-text model on an audio clip and hands back plain text, deliberately separate from add-subtitles-automatically, which burns the same kind of transcript directly onto the video frames instead.
The difference between those two outputs, and what the plain-text path is actually for, is below.
Text you can edit, not a caption track
transcribe_audio returns a written transcript you can copy, edit or feed into another tool first — repurposing a video's dialogue as a blog post, pulling an exact quote from an interview, or drafting a script before a translation or dub pass. add-subtitles-automatically is the other tool for when the goal is captions ON the video; this one is for when the goal is the words, on their own.
Feeds straight into a dub or a translation
A transcript from transcribe_audio can be edited and then used as the script for dub-video or translate-video-to-another-language, rather than transcription being a dead end that has to be manually retyped into a separate translation step.
Where a dedicated transcription service still wins
Sonix's whole product is built around the transcript as the deliverable — export into editor-ready formats and workflows built specifically for podcasters and video producers repurposing long-form audio at volume. Versely's transcribe_audio is a focused utility inside a broader creative app rather than a dedicated transcription workspace, and its own documentation is upfront that accuracy depends on audio clarity — background noise, overlapping speakers and heavy accents reduce it.
Same balance as everything else
Because transcribe_audio runs on the same credit balance as the rest of Versely, a transcript, a dub and a caption pass on the same clip don't require separate subscriptions or an export-then-reimport step between them.
How it works
1. Point to the audio
A video's audio track or a standalone recording, via audio_url.
2. Set the language, or let it auto-detect
Pass a language code explicitly, or leave it to detection.
3. Get the plain-text transcript back
Cartesia's ink-whisper model converts the speech to editable text — no burned-in captions involved.
4. Reuse it
Edit it, translate it, turn it into a dub script, or feed it into a captioning pass if it needs to go back on screen after all.
Where this lives in Versely
Who this fits
- Repurposing a video's dialogue as a blog post or show notes
- Pulling an exact quote from an interview or podcast
- Drafting a script before a translation or dubbing pass
- A searchable, editable record of a video's spoken content
Frequently asked questions
Does this burn captions onto my video?+
No — transcribe_audio returns plain text only. For captions burned onto the video itself, add-subtitles-automatically is the right tool instead.
What speech-to-text engine does it use?+
Cartesia's ink-whisper model.
Can the transcript become a translation or dub script?+
Yes — an edited transcript can feed straight into translate-video-to-another-language or dub-video rather than needing to be retyped.
How accurate is it?+
It depends on the source audio — Versely's own documentation says background noise, overlapping speakers and heavy accents reduce accuracy, the same honest caveat any speech-to-text engine carries.
Other alternatives on Versely
Further reading
Try it inside Versely
The all-in-one AI studio for creators. 60+ models for video, image, voice, music and lipsync in a single app.
Reviewed August 26, 2026. Facts about Sonix on this page are general, publicly known positioning, not pricing or feature claims — see /alternatives for how this page set is scoped. Versely capability links above are pulled from the same live data the rest of versely.studio uses, so they move when the product does.