Make SRT captions from a recording.

#How to make captions from a video

1. Drop your file in the box above. MP4, MOV and WebM videos work, and so do audio files like MP3, M4A and WAV. Up to 10 minutes.
2. Wait while it listens. Your browser pulls the sound out of the file and sends it for speech-to-text in 30-second clips. A 5-minute recording takes about a minute.
3. Check the text. Captions appear in SRT format in an editor. Fix names, product terms and anything the AI misheard.
4. Download an .srt file for YouTube, Vimeo, LinkedIn, Premiere or DaVinci Resolve, or a .vtt file for web players and HTML5 <track> elements.

#SRT or VTT?

Both store the same thing: numbered or unnumbered blocks of text with a start and end time.

• SRT is the most widely accepted. Use it for YouTube, Vimeo, LinkedIn, Facebook, and video editors.
• VTT (WebVTT) is the web standard. Use it with an HTML <video> tag, Video.js, Plyr, or a help center that asks for WebVTT.

If you're not sure, download SRT. Most tools that accept one accept the other.

#Tips for accurate captions

• Speak clearly and close to the microphone. Speech-to-text is very good, but it can't recover words lost under keyboard noise or a fan.
• Fix product names once, then search. If the AI hears "Acne" for "Acme", correct every one in the editor before you download.
• Keep captions short on screen. Two lines of about 40 characters each is easy to read. Split long captions in the editor if you need to.
• Recordings without talking get no captions. The tool captions what's said. If your screen recording is silent, see below.

#Captions for a silent screen recording

Many screen recordings have no voice at all. A caption tool can't caption silence, but Nook can write the narration for you. Upload the recording and Nook reads what happens on screen, writes a line for each step, voices it with an AI voice, and burns in captions timed to the narration. It can also make captions-only videos for reviewers who watch with the sound off.

If you did talk, Nook keeps your own voice, cuts the ums and long pauses, and times the captions to your actual words. Your first video is free.

#Questions

#Is it free?

Yes, with no sign-up and no watermark. There's a fair-use limit per person per hour, which is plenty for a few recordings.

#Is my video uploaded?

No. Your browser decodes the file and sends only the sound, resampled to small 30-second clips, to the speech-to-text model. The clips are transcribed and discarded; nothing is stored. The video itself never leaves your device.

#What languages does it understand?

It detects the spoken language automatically and transcribes in that language. It works well for English, Spanish, French, German, Portuguese, Italian, Dutch, Japanese, Chinese, Korean, Hindi and many more.

#How long can the recording be?

Up to 10 minutes per file. For longer recordings, split the file first, or make the video in Nook, which takes recordings up to 15 minutes.

#Why are some captions slightly off in time?

Timings come from the speech-to-text model and are usually within a fraction of a second. If one caption is early or late, edit its time in the editor before downloading.

#Can I burn the captions into the video?

This tool makes caption files, which players show on top of the video. To get a video with captions built in, styled and timed to a narration, use Nook.