Generate subtitles from any video or audio with Whisper AI in your browser. Hindi and English, edit every line, download SRT, VTT or TXT. Nothing is uploaded.
Who Uses Video to Subtitles
Use it when: you need captions or a transcript for a YouTube video, course lesson, podcast clip or Reel, in Hindi, English or another language, and the file should not be uploaded to a server.
Video to Subtitles Features
- Whisper in your browser: OpenAI's Whisper speech model writes timed subtitles on your device, after a one-time model download.
- Hindi, English and more: Detects the spoken language automatically, or pick it. Hindi comes out in Devanagari.
- Edit before export: Fix words and start and end times line by line, split or merge lines, and preview them on the video.
- Nothing uploaded: Your video never leaves your device. Download SRT, VTT or a TXT transcript.
How to Use Video to Subtitles
Three steps, nothing uploaded.
- Add a video or audio file: MP4, MOV, WebM, MP3, WAV and more.
- Generate subtitles: Pick Tiny or Base and the language, or let it detect.
- Edit and download: Fix lines, then save SRT, VTT or TXT.
Frequently Asked Questions
How do I make subtitles from a video for free?
Drop the video or audio file into the studio, choose the Tiny or Base model and the spoken language (or Detect automatically), and press Generate subtitles. Whisper writes timed lines on your device. Check and edit them in the list, then download SRT, VTT or a TXT transcript. It is free with no sign-up.
Is my video uploaded anywhere?
No. Your browser decodes the soundtrack and the speech model runs on your own device in a background worker. The only downloads are the model files from Hugging Face and the WebAssembly engine, once, and they are cached in this browser.
Does it work for Hindi videos?
Yes. Whisper is multilingual: pick Hindi or let it detect the language, and the subtitles come out in Devanagari. OpenAI notes that Whisper's accuracy varies widely by language, so for Hindi use the Base model and read the lines through before you publish.
Tiny or Base: which model should I pick?
Tiny (about 44 MB) is the fastest and fine for clear English. Base (about 80 MB) is slower and usually makes fewer mistakes, so use it for Hindi, other languages, accents or noisy audio. Each downloads once and is then saved in your browser.
How long does it take?
It depends on your device and the model. Speech recognition runs on your processor, so a computer is much faster than a phone, and Base takes longer than Tiny. The studio shows progress as it goes, and when it finishes it tells you how long the run took. You can cancel at any time and keep the lines already written.
Can I use it on my phone?
Yes, in a modern mobile browser, but it is slower and phones have less memory. Use Tiny, keep the tab open with the screen on, and split long videos into parts with the Video Trimmer first. Files up to 500 MB are accepted.
What is the difference between SRT and VTT?
Both hold the same timed lines. SRT numbers each line and writes times with a comma (00:00:01,000); it is accepted almost everywhere, including YouTube. VTT (WebVTT) is the W3C format web players use, starts with WEBVTT and writes times with a full stop. TXT is just the words, as a transcript.
Are automatic subtitles accurate enough to publish?
Treat them as a strong first draft. W3C's accessibility guidance says automatic captions usually need significant editing, so read the lines through, fix names and numbers, and add important sounds like [music] if the captions are for accessibility. Every line and time can be edited before export.
Related: video to subtitles, video to srt, auto subtitle generator, hindi subtitles, srt file generator free, audio to text with timestamps, video ka subtitle kaise banaye