How speaker diarization works — and where it slips
Splitting a recording by who is talking is a different problem from writing down what they said. Here is what the model actually listens for.
2 min read
The WhoScribe blog
Notes on transcription, speech AI and getting usable text out of audio.
Fast, Balanced, Best: which quality level to pick
Three levels, three different models. What actually changes between them, and when the slowest one earns its wait.
2 min read
Try it
Turn your own recording into text
Paste a link or upload a file and read the transcript in minutes.
- 90+ languages
- Files up to 10 hours
- Web, iOS, Android, Chrome