How to transcribe audio to text
- Drop your audio or video file above, or choose it.
- Pick the language, or leave it on Detect it.
- Press Turn it into text, and wait while the AI listens.
- Copy the text, or save it as a .txt file or .srt subtitles.
How it works
The page runs Whisper, an AI model from OpenAI that learned to hear speech in about 100 languages. It listens in 30-second pieces and writes down what it hears in each one.
It all happens in your browser, so your file is never uploaded. The first run downloads the model, and the next runs start right away.
When your browser lets it, the page runs on your graphics chip, which is much faster. If not, it runs on the processor and takes longer on long files.
Tips for the best text
- Clean up the sound first. Background noise makes the AI guess at words. Run a noisy file through the noise remover before you turn it into text.
- Pick the language if you know it. Detect it only listens to the first 30 seconds. If those are music or another language, set it by hand.
- Check names and numbers. The AI writes what it hears, so it can get a rare name wrong. Read the text through before you share it.
- Use the .srt file for videos. Upload it next to your video on YouTube, or open it in your video editor as subtitles.