# Auto captions that are transcribed on your Mac

Canonical URL: https://getvidiyo.app/features/auto-captions
Last updated: 2026-10-08

> TL;DR: Vidiyo transcribes speech locally with Whisper and turns it into word-timed captions. Pick one of twelve styles, from plain subtitles to word pop and karaoke pills, or ask the assistant to caption the video for you. Captions follow every cut, stay clear of faces, and export burned in or as an SRT file.

Auto captions in Vidiyo start with a transcript made on your own Mac. Every word gets a start and end time, and those timed words become captions you can style, move and correct. Because each caption is tied to the word you spoke, the captions keep up when you cut the video.

## How do I add auto captions to a video on Mac?

Import your footage and either ask the assistant or use the Captions panel. The first time you transcribe, Vidiyo downloads its English speech model once (about 490 MB). After that, transcription runs offline. Captions appear on their own timeline row, already split into readable blocks.

Things you can type:

- "Add bold word-by-word captions."
- "Use the word pop style and keep them above my chin."
- "Captions are covering the product. Move them up."
- "Fix the spelling of Vidiyo everywhere in the captions."
- "Make the captions match the brand colours."

Or do it by hand. The Captions panel has a **Transcript** tab, where you edit words and block timing, and a **Presets** tab, where you pick a look. The common controls stay visible; advanced typography, transforms and fades sit in expandable settings.

## Which caption styles can I use?

There are twelve presets. Each one is a starting point you can restyle.

| Style | What it looks like |
| --- | --- |
| Subtitle, Minimal, Editorial | Calm blocks of text for explainers and longer videos |
| Word highlight, Marker | The spoken word lights up or gets a marker behind it |
| Word pop, Slam, Rise | Words animate in as they're said, for shorts and reels |
| Pill karaoke, Wipe | A pill or wipe follows the speech word by word |
| Bold stack, Typewriter | Big stacked lines or a typed-on reveal |

Captions take an outline colour and width so they stay readable over busy footage. If you set a subtitle style in the Brand panel, new captions follow it.

## How do Vidiyo's auto captions work?

Vidiyo runs whisper.cpp locally, then aligns each word to the audio using Whisper's own timing anchors and measured silences, so words don't stretch into pauses. A word the aligner can't place confidently keeps an estimated time and a warning, so you know which ones to check.

Placement is measured too. When the assistant places captions, Vidiyo computes face-safe positions from the measured face in the frame. After each edit it checks for captions over faces, captions running off the frame, captions hidden under the Reels, TikTok and Shorts interface, blocks that are too long, and speech with no caption at all.

When you cut the video, captions re-resolve through the edit. A trimmed word drops out, and a manual timing fix you made survives a fresh alignment.

## Can I correct a caption word or its timing?

Yes. Open the Transcript tab in the Captions panel and fix the word in place: a brand name, a misheard term, a name the model hasn't heard before. You can also adjust a block's timing there. Corrections are kept by word, so they survive later edits and a refreshed alignment. If the same word is wrong everywhere, ask the assistant ("it's Vidiyo, not video, in every caption") and it fixes every instance in one undoable step. Styling can be scoped too, so you can restyle one block without touching the rest.

## What don't auto captions do?

They aren't a translation tool. Vidiyo captions what was said, in the language it was said. Hindi comes out in Devanagari script, so if you speak Hinglish and want Roman script, you'll fix some words by hand. Word timing is good but not perfect on mumbled or overlapping speech, which is why uncertain words are flagged rather than guessed silently.

Captions are burned into the exported video. For platforms that want a separate file, export an [SRT](https://getvidiyo.app/glossary/srt-file) from the Captions panel.

## Who are auto captions for?

Anyone posting talking-head videos, reels or shorts, where most people watch with the sound off. They're also useful for product demos, where captions make a voiceover easier to follow. If you record in English, it works out of the box. For other languages, download the language pack once from Settings → Transcription.

Pair captions with [silence and filler word removal](https://getvidiyo.app/features/silence-and-filler-word-removal) and you get a tight, captioned cut from one request.

## Honest limits

- Hindi transcribes in Devanagari, so Hinglish captions need some words fixed by hand.
- Word timing is aligned automatically. A word the aligner isn't sure about is marked so you can correct it.
- Export is 1080p at 30 fps, and timelines max out at 5 minutes.

## Frequently asked questions

### Does Vidiyo add captions automatically?

Yes. Vidiyo transcribes the audio on your Mac and generates word-timed captions from it. You can ask the assistant to "add bold captions" or generate them from the Captions panel, then restyle or correct any word.

### Is caption transcription done offline?

Yes. Transcription runs on your Mac with whisper.cpp, and the audio never leaves it. The English model (about 490 MB) downloads once on first use and then works offline.

### Can Vidiyo caption languages other than English?

Yes, with the downloadable language pack (about 575 MB), which covers all the languages Whisper supports and can detect the language for you. Hindi comes out in Devanagari script, so Hinglish captions need some words fixed by hand.

### Do captions stay in sync after I cut the video?

Yes. Each caption word is tied to a transcript word, so trims, splits and deleted ranges keep the captions on the right words. Words cut from the video drop out of the captions.

### Can I export captions as an SRT file?

Yes. Captions are burned into the exported MP4, and you can also export them as an SRT file or import an SRT you already have. Imported captions stay editable.

## Related

- [Edit video by chatting](https://getvidiyo.app/features/chat-video-editing)
- [Remove silences and filler words automatically](https://getvidiyo.app/features/silence-and-filler-word-removal)
- [Word-level captions](https://getvidiyo.app/glossary/word-level-captions)
- [Burned-in captions](https://getvidiyo.app/glossary/burned-in-captions)
- [SRT file](https://getvidiyo.app/glossary/srt-file)
- [How to make word-by-word captions](https://getvidiyo.app/blog/word-by-word-captions)

Request access to the Vidiyo beta: https://getvidiyo.app/request