Frequently Asked Questions
Is my video really not uploaded?
Correct. Both ffmpeg.wasm and Whisper run entirely in your browser tab on your CPU. Your media file stays on your device.
Why does it download a 75 MB file the first time?
That's the Whisper model weights (downloaded once, then cached forever). Bigger models are larger — base is ~150 MB, small is ~480 MB.
Which model size should I choose?
Tiny is fast and accurate enough for clear English. Base is a good balance. Small is best for noisy audio or non-English. Bigger = slower.
What languages are supported?
Whisper supports 99 languages. The English-only models (tiny.en, base.en) are faster and more accurate for English. Use the multilingual models for any other language.
How long does transcription take?
Roughly real-time on a modern desktop CPU with Whisper-tiny. So a 5-minute video takes about 5 minutes. Bigger models are 2–5× slower.
Can I edit the subtitles after?
Yes. Click any line in the editor below the result to edit text or adjust timestamps before exporting.
Can I burn the subtitles into the video?
Yes. After transcription, click 'Burn into video' to render an MP4 with subtitles overlaid. Choose from classic or TikTok-style.
Output formats?
SRT, VTT, plain text, and JSON (with timestamps).