Formats
Supported media format guidance
In browser transcription, “supported format” really means “supported by the browser decoding stack plus the surrounding media tooling.” Practical compatibility matters more than a long marketing list.
Typical workflow expectation
Cowslator is most comfortable with common audio and video formats that browsers already handle well or that can be normalized through the local processing pipeline. If a file plays reliably in a modern browser, it is a good candidate for transcription.
Audio
Batch selection recognizes WAV/WAVE, MP3, OGG/OGX, M4A, FLAC, WMA, AAC, ALAC, AIFF/AIF, Opus, AMR, APE, AC3 and OGA. Files are normalized to mono 16 kHz WAV with the local FFmpeg runtime before transcription.
Video
MP4, WebM, MKV, MOV, AVI, MPEG, MPG, M4V, 3GP and WMV files can enter the same batch workflow; FFmpeg extracts their audio track for speech transcription.
Folders
Batch selection filters by a supported media extension, skips documents and other files, and processes the remaining files in order. Browser memory and FFmpeg codec support still limit which files can be decoded.
When preprocessing helps
If source media is unusually large, oddly encoded, or part of a messy archive, it can help to normalize it before transcription. Creator workflows often benefit from extracting clean audio or trimming long captures into more manageable segments before the speech model runs.