Subtitle Generator

Generate subtitles for your video automatically - free, private, no upload.

Generate subtitles for a video without installing anything or sending your file anywhere. This subtitle generator runs Whisper speech recognition locally in your browser by default - free, with no upload and no sign-up - and previews the captions live over your video before you export them.

Drop a video or audio file here, or click to choose one

MP4, MOV, MP3, WAV and more - up to 30 minutes

Runs locally with Whisper speech recognition. Your file is never uploaded.

Why use this subtitle generator

Runs in your browser

Your video is processed locally, in a Web Worker. Nothing is uploaded, and there's no server in the loop.

See captions on your video

Captions play back live over the video as it generates them, so you can judge timing and wording immediately, not after export.

Edit before you export

Fix names and technical terms directly in each caption, before you download anything.

SRT and VTT export

Download subtitle files ready to upload to YouTube, drop into a video editor, or open in the subtitle editor for finer sync work.

How to generate subtitles for a video

Drop a video or audio file above, or click to choose one.
Choose a model size and language, or leave both on the defaults.
Click Generate captions. The first time you do this, your browser downloads the speech model - a one-time download, cached afterwards.
Watch captions play back live over your video as it generates them.
Click into any caption to fix names or wording, or click a timestamp to jump the video to that moment.
Download the result as .srt or .vtt.

Common subtitle generator problems and fixes

The first run is slow: that's the one-time model download, not a bug. It's cached afterwards, so every later run on this device starts immediately.
Accuracy is poor on accents, crosstalk or background noise: this is a known limit of speech models generally, not just this one. Try the Best quality model size, or edit the captions directly.
Names and technical terms come out wrong: click into the caption and fix it in place - your edit is used in the preview and every export.
Long files fail or crash the tab: split the video into parts under the 30-minute cap and generate captions for each one separately.
No captions come back at all: check the file actually has sound - a silent track, the wrong audio track in a video file, or a muted recording are the usual causes. The tool reports "no speech detected" rather than returning an empty result.
Captions look out of sync in another player after export: re-check the timing in the subtitle editor, which can shift or scale the whole file against the real video if the drift is consistent.
It won't run in this browser: the local engine needs WebAssembly, and runs faster with WebGPU. Update to the latest Chrome, Edge, Firefox or Safari.

How this compares to Kapwing, VEED and YouTube's auto-captions

Hosted tools like Kapwing, VEED and YouTube's own auto-caption feature can burn captions directly into the video and offer style presets, but they require uploading your file to their servers (YouTube specifically needs the video published or unlisted on the platform first). This tool trades that for privacy: your video never leaves your device, there's no account, and no size-based paywall - the tradeoff is that it only generates and previews subtitle files, it doesn't produce a new video with captions baked into the picture.

Just want a plain transcript instead of timed captions? Use audio to text. Already have an SRT or VTT file and need to fix its timing instead of generating one from scratch? Open it in the subtitle editor. You can also browse all Meetrix features, or read how Meetrix's self-hosted Jitsi platform handles closed captions and live transcription.

Frequently Asked Questions

Is this subtitle generator free?

Yes. It's free to use, with no sign-up, no watermark, and no limit on the number of files - just the per-file length cap listed above.

Is my video uploaded anywhere?

No. Your file is decoded and transcribed entirely in your browser, in a background worker, using Whisper speech recognition. It never leaves your device or gets sent to Meetrix or anyone else.

How accurate are the generated subtitles?

On clear, single-speaker audio it does well. Accuracy drops on strong accents, overlapping speakers, background noise, and technical jargon or names - that's true of any speech model, not just this one. Click into any caption to fix what it gets wrong before exporting.

Can I edit the captions before exporting?

Yes. Click into any caption in the list and type - your edit is used immediately in both the live preview and every export.

What file formats can I export?

SRT and VTT, the two subtitle formats almost every video platform and player accepts.

Which languages are supported?

Whisper supports dozens of languages. Leave the language on Auto-detect, or pick one explicitly from the dropdown if auto-detection guesses wrong on a short or noisy clip.

How long can my video be?

Up to 30 minutes - that's the point at which a browser tab reliably starts running low on memory for this kind of processing. Longer files: split them into parts and generate captions for each one.

Does this work with audio-only files?

Yes. Captions still generate and export normally as SRT/VTT - there's just no video to show the live caption preview over, since there's no video track to overlay it on.

Can this burn the captions permanently into the video?

No, not this tool - it generates and previews captions and exports SRT/VTT subtitle files, but doesn't re-encode a new video file with captions baked into the picture. Most video editors and platforms (YouTube, Premiere, CapCut) can burn in an SRT/VTT file you upload from here.

Can it tell different speakers apart?

Not yet. Speaker labels aren't in this version - captions come back as one continuous stream, split into timed segments rather than by who's talking.

Why is the first run slow?

The first time you generate captions, your browser downloads the speech model (40 MB to a few hundred MB, depending on the size you pick). That download is cached, so every run after that starts immediately.

Which browsers are supported?

Any recent Chrome, Edge, Firefox or Safari. Chrome and Edge can use your GPU (WebGPU) and run faster; other browsers fall back to WebAssembly, which works everywhere but is slower.

Does this work on a phone?

Yes, but expect it to be slow. Both the model download and caption generation take longer on a phone than on a laptop - it works, it just needs patience.

How is this different from the audio to text and subtitle editor tools?

This tool is built specifically for captions: it previews them live over your video and exports straight to SRT/VTT. Audio to text is for when you want a plain transcript (.txt/.md) rather than timed captions. Subtitle editor is for when you already have an SRT/VTT file and need to fix its timing or convert its format - open this tool's export there for finer sync control.

Captioning at volume? Meetrix builds transcription and captioning pipelines.

For teams captioning many videos, see our self-hosted Transcriber AMI - deployed on your own AWS account, no per-file upload to a third party.

Talk to Us