Demo · Hamsa STT
From audio to transcript.
Drop in a recording (a call, a meeting, a voice note) and read it back. Turn on diarization to see who said what.
Drop your audio here.
or click to browse · MP3 · WAV · FLAC · M4A · OGG · up to 20 MB
Files are processed for this demo only · diarization audio auto-deletes within 48h
01
Hamsa STT
The audio streams through our proprietary speech model, built for real-life Arabic, dialects and code-switching.
02
Speaker separation
In parallel, not after, diarization separates the voices and timestamps every speaker turn, so you wait for the slower job only.
03
Overlap merge
Each transcribed segment is assigned to the speaker it overlaps most: the same alignment used in production pipelines.
Need this at scale, on-prem or in-region?
Talk to us →


