Support
How can we help?
Kayit runs entirely on your Mac, so most questions are about getting set up and getting the most out of it. Browse the FAQ below, or send us a message and we'll get back to you.
Frequently asked questions
Does Kayit send my audio or transcripts to the cloud?
No. Kayit transcribes entirely on your Mac using on-device machine learning. Your audio, transcripts, and quotes never leave your machine and are never uploaded to any server, account, or third-party API.
Do I need an account or sign-in to use Kayit?
No. Kayit requires no account and no sign-in. There is nothing to log into and no personal information to provide.
What do I need to run Kayit?
Kayit runs on a Mac with Apple chip (M1 or later) running macOS 15 (Sequoia) or later. An internet connection is required only on first launch to download the transcription model.
Why does Kayit need the internet on first launch?
On first launch Kayit downloads its speech-recognition model (roughly 800 MB) over HTTPS from Hugging Face. This is machine-learning model data, not your audio. After the one-time download, transcription works fully offline.
Does Kayit identify who is speaking?
Yes. Kayit performs automatic speaker separation on-device, labelling turns as "Speaker 1", "Speaker 2", and so on. You can rename each speaker, and the renaming applies across the whole transcript.
What formats can I export?
You can export a transcript as clean Markdown, plain text, or a quotes-only file. Kayit also writes a .transcript.json sidecar next to your audio so your work is portable and future-proof.
How is Kayit different from Otter.ai, Descript, Rev, or Trint?
Those tools upload your audio to their servers to transcribe it. Kayit is built to never do that — all processing happens locally on your Mac, which protects confidential sources and interview material by design.
How much does Kayit cost?
Kayit starts with a free trial. After that it is $9.99 per month, or $99 per year. One subscription covers the app on your Mac — with no per-minute charges, no upload limits, and no ads.
Troubleshooting quick hits
- Stuck on first launch? The initial model download (~800 MB) needs a stable internet connection. Keep Kayit open and it will resume; after that, it works offline.
- Transcription seems slow? The first run also warms up the on-device model. Later transcriptions are faster. Closing other heavy apps frees up the Neural Engine and memory.
- Speakers not separating well? On the setup screen, set the exact number of speakers instead of Auto-detect for cleaner results, then use the Speakers panel to rename them.
- Want to protect a recording? Keep your library folder out of iCloud, Dropbox, or other syncing folders — Kayit will warn you if it detects one.