Audio to Text & Captions
Transcribe short English PCM WAV recordings locally into editable TXT, SRT and WebVTT drafts.
1. Choose a short recording
English speech · 16-bit PCM mono/stereo · 6 MB · 30 seconds maximum. Use the WAV trimmer for a shorter clip.
2. Generate a local draft
First use downloads approximately 70 MB of assets. A desktop browser with sufficient memory is recommended. Recognition may omit or invent words, especially with noise, music or accents. Nothing is uploaded.
3. Review words and timing
Your editable draft appears here after recognition.
Practical example
Choose a clearly spoken ten-second English WAV. Listen to the original, correct the transcript and caption times, then save TXT or captions.
Limits · supported files
16-bit PCM WAV mono/stereo at 8–96 kHz, up to 6 MB and 0.1–30 seconds. Approximately 70 MB first-use model/runtime download; five-minute processing budget. Desktop devices with sufficient memory recommended. English only; no MP3/video, diarization, translation or accuracy guarantee. Model estimates may omit or invent words. Caption times can be inaccurate or unavailable. Transcript and caption edits are independent; verify both against the recording. Inputs remain on-device.
Step-by-step guide
How to use Audio to Text & CaptionsReview the example, result checks and limitations before using your own inputs.