This tool requires the Media tier. The engine is already loaded — sign in and upgrade to run it on your audio.
Free is enough for most one-off jobs. Pro raises the file and batch caps; Pro + Media unlocks GB-scale streaming and unlimited duration.
Larger files supported on Developer (5 GB CSV) and Enterprise (unlimited). All processing happens in your browser — files never reach a server.
Drop your audio file (any common format)
Pick language and output format (plain text, SRT, VTT, or JSON segments)
Download the transcript
0 bytes uploaded. Audio Transcription runs entirely in your browser using FFmpeg.wasm 8.1, RNNoise, and the Web Audio API. Your audio stays on your device at all times. No data is sent to any server — critical for sensitive interviews and confidential calls.
No. Transcription runs on the paired runner using a local Whisper-class model. Audio bytes never reach JAD servers.
100+ languages including English, Spanish, French, German, Mandarin, Japanese, Hindi, Portuguese, and Arabic. Set the language hint for best accuracy or leave on auto-detect.
Yes — pick the JSON output format. Each segment includes start/end timestamps, and the model can emit per-word offsets for fine-grained alignment.
Cut start/end timestamps from any audio file. JAD re-encodes once at the source quality — your file never leaves the browser.
Open toolDetect and cut silences below a threshold (default -40 dB, 0.5 s minimum) from any recording. Tighten interviews and solo podcasts in seconds.
Open toolTwo-pass EBU R128 loudness normalisation in your browser. Hit -16 LUFS for podcasts, -14 for Spotify/YouTube — true-peak limited, FFmpeg 8.1 engine, no upload.
Open tool