Local AI Audio Transcriber
Turn a short English recording into editable text using the open-source Whisper model on your device.
Transcribe a short recording
Supported formats depend on your browser. Clear speech works best. No microphone access is requested.
Ready. No model downloads until you start.
Uses Whisper tiny.en via Transformers.js, running on CPU. Transcription can take longer than the recording on slower devices.
Your result
Speech models can mishear words or invent text during silence. Review the transcript while listening to your recording.
Audio to useful notes
Choose a file, approve the first download and start transcription. The browser decodes your recording and runs Whisper tiny.en locally. No speaker identification, translated transcript or guaranteed timestamp accuracy is claimed.
Summarize your transcript · Convert existing subtitles · All AI tools