OpenShortsVizard AI video to text: what you get, and what it costs
Vizard's "video to text" is a transcript with timestamps, produced from the same pass that finds the clips and burns the captions. It is part of the entry plan, which starts at $19.99/month, and the free plan allows 120 upload minutes and 10 exports.
A transcript is a by-product of the transcription stage every clipper already runs, which is why no tool charges for it separately and why it is not worth choosing a tool over.
OpenShorts transcribes with faster-whisper at word level and returns the transcript alongside the clips, subtitles included, free when self-hosted and from $12/month hosted.
Word-level timestamps, burned-in captions and the clip list, from one pasted link. 20 free minutes a month.
What "video to text" means in a clipping tool
Three different outputs get described as video to text, and they are not interchangeable. A transcript is the words with timestamps. A subtitle file is the same words shaped for playback, cut into readable lines. A summary or article is a written document generated from the words, which is a language-model task sitting on top of the transcript rather than a transcription task at all.
When a clipping tool advertises video to text, it means the first two. They fall out of the transcription pass the tool has to run anyway to find the good moments, which is why they are bundled rather than sold.
What Vizard gives you
Vizard runs the whole thing in the browser and treats the timeline as the product: it transcribes the upload, lets you edit the captions and the clip boundaries on a timeline, and exports both the clips and the text. Multi-language subtitles are one of its stronger features. Its entry plan starts at $19.99/month, and the free plan allows 120 upload minutes and 10 exports.
Vizard is built around a browser editing timeline, so it sits between a pure auto-clipper and a full editor. That is useful if you intend to hand-adjust every clip, and overhead if you do not.
Doing the same thing with OpenShorts
OpenShorts transcribes with faster-whisper and keeps a timestamp for every word, not every sentence. Word-level timing is what makes the subtitle file land on the right frame and what lets the clip boundaries fall inside a sentence instead of at the nearest one. The transcript is returned with the finished job next to the clips, and the burned-in captions are cut from it.
- Self-hosted: free, MIT-licensed, no watermark, no cap. Transcription runs locally, so the audio never leaves your machine.
- Hosted: 20 free minutes a month with no credit card, then flat plans from $12/month with no watermark.
- Via API or MCP: the transcript and the clips come back from the same job, so an agent can summarise the text without a second transcription bill.
Which one should you pick?
If you want to hand-correct captions on a timeline before exporting, Vizard's editor is the better fit and it is not close. If you want the text as a by-product of clipping at volume, or you need the transcript to stay on your own machine, self-hosted OpenShorts is free and the transcript comes with the job.
Common questions
- Does Vizard AI convert video to text?
- Yes. It transcribes the video it processes and gives you the text with timestamps alongside the clips and subtitles. Its entry plan starts at $19.99/month, and the free plan covers 120 upload minutes and 10 exports.
- How do I get a free transcript from a video?
- Self-host OpenShorts: transcription runs locally with faster-whisper at word level and the transcript comes back with the finished job, at no cost and with no watermark. The self-hosted edition needs Docker and a Google Gemini API key for the moment scoring, whose free tier covers 1,500 requests a day.
- Is word-level timing important in a transcript?
- For subtitles, yes. Sentence-level timestamps force captions to appear for the whole sentence at once, which is where the timing drifts out of sync with the speech. Word-level timestamps let each caption start and end on the word being spoken, and they are also what lets a clip boundary land mid-sentence without cutting a word in half.
Sources
- Vizard plan limits checked 2026-07-27 on the vendor's public pricing page.
- OpenShorts transcription and subtitle stages in the project source at github.com/mutonby/openshorts.