Turn any recording into the 16 kHz mono file speech models read, as WAV or a small MP3. Runs in your browser; the file never leaves your device.
Three steps. Zero uploads. The entire pipeline runs in your browser tab.
Click the drop zone or drag a file in. We read it locally — no network call is made to open it.
FFmpeg downmixes to mono and resamples to 16 kHz inside a Web Worker via WebAssembly. No server round-trip.
Grab the new file with one click. Everything stays local until you choose where it goes.
Most speech models process 16 kHz mono internally. Converting first removes resampling surprises and cuts file size versus 44.1/48 kHz stereo.
Not for speech recognition — speech intelligibility lives below 8 kHz (the Nyquist limit of 16 kHz sampling), which is why telephony and ASR standardized there.
No. WhipScribe converts every recording to 16 kHz mono on its own servers, so upload the original. This tool is for your own speech-recognition experiments, or to make a file small enough to send.
No — conversion runs via ffmpeg-wasm in your browser.
The first click loads a 32 MB WebAssembly build of FFmpeg. Your browser caches it, so later runs in the same tab start at once.
50 MB per file, because the work happens in your browser's memory. A longer recording does not need this step to be transcribed: WhipScribe takes one file up to 10 hours or 5 GB as it is.
Paste your YouTube / podcast / Dropbox URL in Whipscribe — we transcribe and let you chat with the transcript. Free preview of the opening minutes; packs from $4.
Open Whipscribe →Cut a clip by entering start + end times. Same format in, same format out.
Video → audioPull the audio track out of an MP4 / MOV / MKV as MP3 or M4A.
TranscriptionUpload a file or paste a URL — our server transcribes and returns a searchable transcript.
SubtitlesTurn a transcript into a standard .srt or .vtt subtitle file.
Drop the same file here and get a timestamped transcript — instant preview, no signup.
Once the file is ready, send it as it is: see MP3 to text for long recordings, MP3 to text and transcribing a podcast episode.