Auto captions for any video, free

Auto captions are written by a speech recognition model instead of a person. Ours is Whisper large-v3 turbo. It listens to the audio, picks the most likely words, and marks when each one starts and ends. We then group the words into lines of up to about 42 characters and roughly 5 seconds. Vertical videos get shorter lines, around 22 characters, so they fit a phone screen.

With one clear voice and a decent mic, most lines come out right the first time. The mistakes show up in the same places again and again:

• Names of people, brands and products. The model spells what it hears, so "Groq" can turn into "grok".
• Heavy accents and fast talkers, worse still with music underneath.
• Crosstalk. When two people talk at once, it follows one voice or mashes both into a line nobody said.
• Numbers. "Fifteen" can come back as "50", or as a word in one line and digits in the next.

So proofread before you export. On a short clip it takes a minute or two: click a line and the video jumps to that moment, so you can hear what was actually said. YouTube gives creators the same advice about its own automatic captions.

If someone talks in it and it runs 30 minutes or less, you can caption it here.

• TikToks, Reels and Shorts. Karaoke, Classic and Boxed sit low in the frame, where those apps put their buttons and post text. Drag the position slider up, or use Bold pop, which sits higher.
• YouTube videos, tutorials and course lessons. Here a subtitle file often beats burned-in text. The SRT generator opens straight on the file downloads.
• Podcast clips, interviews and Zoom recordings. If you only need the words, convert the video to text and copy the transcript.
• Videos in Spanish, Hindi, Portuguese, Arabic, French, German, Japanese and a few dozen other languages.

The highlighted-word look has its own guide: word-by-word captions, how the timing is done and when the style is worth it.

Sometimes you should. YouTube, TikTok and Instagram all make captions for free, and if a video only ever lives on one app, theirs may be enough. The real difference is where the captions end up.

YouTube adds automatic captions after upload as closed captions that viewers switch on, and you can correct them in YouTube Studio. Our guide to YouTube's auto captions covers turning them on, off and fixing them.

TikTok transcribes your speech when you tap Captions on the preview screen, and you can edit each line before posting. They're tied to that one post, so the same clip on Reels starts from nothing. Our steps for adding captions on TikTok cover both routes.

Instagram can switch on automatic closed captions for Reels, and its Captions sticker puts text on the video in the app. The Reels captions guide walks through both.

Captions burned in here are part of the picture. They show with the sound off, and they stay put when you post the same file somewhere else. That's the open-caption approach, and the guide to open and closed captions explains the trade-off. Plenty of people do both: burned-in captions for social, an SRT for YouTube.