Workflow
One path, kept short.
Geaso is not a chat assistant. It focuses on the one thing you repeat all day: turn speech into text and put that text where you need it.
Trigger
Use the global hotkey, floating panel, or menu bar entry when you want Geaso to listen.
Transcribe
Geaso streams audio to the speech backend and shows partial text while the session is running.
Review
The final transcript stays visible before sending, so it never disappears at the wrong moment.
Insert
Geaso tries to paste into the app that was active before recording, with copy fallback if insertion fails.
Product
Built for repeat use.
The interface stays compact, diagnostic, and native to macOS instead of turning dictation into another oversized workspace.
Hotkey and floating panel
Start from the keyboard or keep a small speech surface near your current work.
Live partial text
Recording, processing, draft, and error states stay visible so you know what is happening.
Text does not vanish
If automatic paste is unavailable, the transcript remains available for copy and troubleshooting.
Remote and Apple speech
Remote uses Geaso's speech backend. Apple mode keeps a local macOS Speech path available.
Permission checks
Microphone, speech, accessibility, service, and quota states are separated for clearer fixes.
Local session trace
Recent recordings can keep audio, transcript, task IDs, outcomes, and diagnostics on your Mac.
Privacy
Clear boundaries for voice data.
Geaso handles microphone audio, so the privacy rules are explicit: record only when triggered, keep diagnostic retention short, and do not train models on your recordings or transcripts.
Only when you record
Geaso does not keep the microphone open in the background. Recording begins from an intentional user action.
Short diagnostic retention
Server diagnostics keep metadata and necessary transcript text for up to 7 days. Audio is kept only when you explicitly send it for support.
No model training
Your audio and transcript content are not sold, used for ads, or used to train models.
TestFlight beta is opening soon.
The first beta will use Geaso's cloud speech backend. Read the privacy policy first if you want the data flow in plain language.