Support

Questions and problems: seduta@spert.ai. The answers below cover what most people ask.

Getting started

The first transcription needs the speech model. Seduta downloads it once (about 470 MB). On the iPhone it waits for Wi-Fi unless you allow cellular downloads in Settings. Until then the home screen shows a Download button.

Recording a call on the Mac. Leave “Ask to record calls” on. When Zoom, Teams, Meet or another app opens the microphone, a notification offers to record. Or press Start Recording yourself. Keep Seduta open at login (Settings › General) so it is there when the call starts.

Recording a call on the iPhone. iOS gives phone calls exclusive use of the microphone, so Seduta records in-person meetings. Record a call in the Phone app; when it ends, share the recording from Notes to Seduta and it becomes a meeting.

Summaries

“Turn on Apple Intelligence”: on-device summaries need Apple Intelligence on, in System Settings on a Mac with Apple silicon or an iPhone 15 Pro or later with iOS 26. Otherwise use an endpoint.
“The endpoint refused the API key”: check the key in Settings › Summaries. EUrouter keys start with a dashboard at eurouter.ai; the Settings screen shows what a key has spent.
Your own server: any OpenAI-compatible endpoint works, including Ollama on your Mac at http://localhost:11434/v1. Lower “Words per request” for small models.

Speakers and names

Your own voice is “Me”. Other people start as “Speaker 1, 2, …”. Tap a name in the transcript to rename it; Seduta remembers the voice for next time.
Names and terms the transcriber gets wrong go in the names-and-terms file (Settings › Storage). Names from calendar invites are added by themselves.

Files

Every meeting is a folder in iCloud Drive › Seduta: meta.json, transcript.md, summary.md, and the recording until it is removed. Deleted recordings sit in “Recently Deleted” for a few days. You can open, copy or delete anything there; Seduta only reads what it finds.

Seduta 1.0. Spert, Switzerland.