How do automatic subtitles work?
Subtitling a video by hand takes hours: you have to listen, type and time every sentence. This generator does it in minutes with Whisper, OpenAI’s open-source speech recognition model and one of the most accurate available. It understands almost a hundred languages, a wide range of accents and technical vocabulary, and adds the punctuation for you.
The tool extracts the audio from the video, splits it into chunks at natural pauses so no words get cut off, and transcribes each one with its timestamps. It then arranges the text into subtitles that follow readability guidelines: at most two lines of 42 characters (or 32 for vertical videos), with no overlaps and just the right time on screen.
When it’s done, you see the video with the subtitles in place and the full list of lines. Click any line to fix a word or a name, and click its time to jump to that moment. Download the result as SRT (the universal format for YouTube, Premiere, DaVinci Resolve or CapCut), as VTT for the web or as plain text. If you need the subtitles in another language, you can translate them into English directly or send them to the subtitle translator.
The AI runs in your browser: your video is never uploaded to a server. The free trial lets you subtitle one video; with ToolStudio Pro you get unlimited use and can generate subtitles for many videos at once.
How to generate subtitles for a video
- Drag in your video or click to choose it.
- Leave the language on “Detect automatically” or pick it, and choose the accuracy.
- Click “Generate subtitles”. The first time, the AI model is downloaded; after that it stays saved in your browser.
- Review the subtitles over the video and fix any word by clicking it.
- Download the SRT or VTT file and upload it to YouTube, your social network or your video editor.
What are automatic subtitles for?
- Subtitle YouTube videos to reach more people and improve their search ranking.
- Add captions to Reels, TikToks and Shorts, which most people watch with the sound off.
- Make your videos accessible to people who are deaf or hard of hearing.
- Subtitle courses, webinars and internal training.
- Create the base SRT for translating a video into other languages.
- Save hours when editing interviews, documentaries and video podcasts.
Which accuracy should you choose?
- Fast: a lightweight model, ideal for clear speech and short videos.
- High: more accurate with accents, background noise and technical terms. It’s the default on computers with a compatible graphics card.
- Maximum: Whisper large-v3 turbo, the best accuracy. Needs WebGPU (Chrome or Edge on a computer with a graphics card) and a large download the first time.
Supported formats
- Input: MP4, MOV, WebM and MKV videos, and MP3, WAV, M4A, AAC, OGG, OPUS and FLAC audio. Up to 2 GB per file.
- Output: SRT and WebVTT (VTT) subtitles and plain TXT text, with or without timestamps.
- Languages: English, Spanish, French, German, Italian, Portuguese, Catalan, Galician, Basque and almost a hundred more, with automatic detection.
Frequently Asked Questions
How accurate are automatic subtitles?
With clear speech, Whisper gets the vast majority of words right, but no AI is perfect: names, brands or very specific terms may need correcting. That’s why you can edit every subtitle before downloading.
What languages does it support?
Almost a hundred languages, including English, Spanish, French, German, Italian, Portuguese, Catalan, Galician and Basque. The language is detected automatically, although choosing it makes the result a little more reliable.
Can I generate the subtitles in another language?
Whisper can translate directly into English: choose “Translated to English” under “Output”. For any other language, download the SRT and use our subtitle translator.
Why is it slow the first time?
The first time, your browser downloads the AI model. After that it stays saved and starts right away. Speed depends on your device: on a computer with a graphics card running Chrome or Edge, it’s several times faster.
Is my video uploaded to a server?
No. The AI (Whisper, OpenAI’s open-source speech recognition model) is downloaded to your browser and runs on your own device. Your audio is never sent to a server; the only thing downloaded from the internet is the model, and only once.
Is it free?
You get one free try per tool: one file. You can reprocess that same file with different settings as many times as you like, and if something fails, your free try isn’t used up. To process new files you need ToolStudio Pro, which gives you unlimited use of every tool, batch processing (several files at once, downloaded together in a ZIP) and no ads. You can cancel anytime.
