Subtitle Generator
Generate SRT or VTT subtitles from MP4, MOV, or WebM video with speaker labels, then edit any line before export. Free, English speech, nothing uploaded.
Your audio stays inside this browser tab, never reaches a server, and nothing is saved to your computer. The first visit downloads the AI models into your browser cache, about 210 MB once (300 MB in compatibility mode). After that they load from the cache.
Pick an audio or video file to transcribe.
Record from your mic, capture another browser tab, or upload an audio or video file. Click any "Speaker 1" label to rename it, and click any line's text to fix the wording before exporting. Clearing a line deletes it.
What is the Subtitle Generator?
A subtitle generator listens to the speech in a video and turns it into timed caption files. This one runs the Whisper speech-recognition model directly in your browser using WebGPU or WebAssembly: it reads the audio track of your MP4, MOV, or WebM, transcribes it with timestamps, tells speakers apart automatically, and exports standard SRT or VTT files that YouTube, Premiere, DaVinci Resolve, VLC, and every major player accept. Because the model runs on your device, the footage itself is never uploaded anywhere.
How to use the Subtitle Generator
- 1
Upload a video or audio file
Pick an MP4, MOV, WebM, MP3, or WAV. The first visit downloads the AI models into your browser cache (about 210 MB, or 300 MB without WebGPU), and later runs skip that download.
- 2
Let the AI transcribe
The audio track is transcribed with timestamps, and speakers are detected and labeled automatically. A progress bar and the latest words show how far it has got, and Cancel stops it at any point.
- 3
Edit lines and name speakers
Click any line to fix the wording, clear a line to delete it, and click a speaker label to replace "Speaker 1" with a real name.
- 4
Export SRT or VTT
Download the finished captions as SRT for video editors and YouTube, VTT for web players, or plain TXT for a readable transcript.
Frequently Asked Questions
Which video formats can it subtitle?
Anything your browser can play: MP4, MOV, and WebM cover nearly all real-world files. The tool reads the file's audio track, so video resolution and codecs don't matter as long as the audio decodes.
Is my video uploaded to a server?
No. The speech-recognition model runs inside your browser via WebGPU or WebAssembly. You can open DevTools and watch the Network tab while it transcribes: after the one-time model download, no data leaves your machine.
What's the difference between SRT and VTT?
SRT is the older, universally supported format that video editors and YouTube expect. VTT is the web-native format used by HTML5 video players and supports speaker voice tags. This tool exports both, cut into short cues of up to two lines, with speaker names when the recording has more than one voice.
Can I edit the subtitles before exporting?
Yes. Click any line's text to correct it, clear a line entirely to delete it, and rename detected speakers to real names. Your edits flow straight into the exported SRT and VTT files.
How accurate is it, and which languages work?
It uses Whisper, the same model family behind most commercial transcription tools, and it handles English only for now. Speech in other languages comes out garbled. Heavy accents, crosstalk, and loud background music reduce accuracy, which is what the built-in editor is for.
Is there a length or file-size limit?
Browsers refuse to decode more than about 90 minutes of audio in one go, so split longer videos into parts. Processing time scales with the length and your hardware. A 30 minute recording took about 3.5 minutes on a desktop PC with a dedicated graphics card and about 9 minutes in a browser without WebGPU. Long files also use a lot of memory, so on a laptop with little RAM the tab may close.