Roadmap
Three states, and nothing promised to a day. Shipped items are in the release notes; the rest says what it is for, not when it lands. If something you need is in the last column, writing to us is what moves it.
Shipped
Two windows, four switches
Transcript and prompter as separate windows, each function switchable on its own and mid-call, from the menu bar or the window itself.
Recognition on your own machine
Whisper models download inside the app; with local recognition the audio of the call never leaves the Mac.
Summary, tasks and export
A call becomes a document: transcript, what was agreed, the tasks that follow, exported with or without the translation.
Building now
Windows build
The same three functions on Windows. It captures audio and shows the windows today; what is not finished is the reliability of capture when audio devices change, which is exactly the part that must not be shipped half-done.
Sign in with Google and Apple
Both flows are written and waiting on the provider registrations. Email codes keep working either way.
Considering
Cloud sync of history
History is local by design. Sync would be opt-in, per device, and encrypted — worth doing only if it can be done without weakening that default.
iPhone and iPad
The interesting case is a phone call rather than a video call. It depends on what the platform allows, so it stays a question until a prototype answers it.
Team accounts
Shared presets and glossaries, one invoice, per-seat minutes. Asked for by support and sales teams; not started.