Is my audio uploaded anywhere?
No. Transcription runs on an AI speech model downloaded to your browser and executed on your own device. Your audio file is never sent to a server.
Why does it need to download something first?
The first time you use this tool, your browser downloads a small speech recognition model (about 40-80MB). Your browser caches it, so it's instant on future visits. This is what lets the tool work without a paid server API.
How accurate is it?
It uses a compact version of OpenAI's Whisper model. Accuracy is good for clear speech in many languages, but can drop with heavy background noise, overlapping speakers, or strong accents.
Is there a file size or length limit?
No hard limit, but longer files take longer to process since everything runs on your device's CPU or GPU. A few minutes of audio typically finishes in well under a minute on a modern laptop.
Is this free?
Yes, completely free with unlimited use.