If the first attempt doesn't work.

Two usual causes. First, the cursor: Acousmos types at the insertion point, so click into a text field before dictating. Second, Accessibility permission: without it the transcription completes but can't be typed anywhere. Check System Settings → Privacy & Security → Accessibility, toggle it off and on if needed, and relaunch the app.

Confirm Microphone permission in System Settings → Privacy & Security → Microphone. Then open System Settings → Sound → Input and watch the level meter while you speak — if it doesn't move, the selected input device isn't receiving sound. External microphones and some headsets grab the input role when they connect; pick the device you actually want to speak into.

Open Preferences → Hotkeys and confirm the style you're using has a hotkey assigned and enabled. If you use a bare key like Fn, Right Command, or Right Option as a push-to-talk trigger, approve Input Monitoring when macOS asks. And if another app owns the same global shortcut, one of them will lose — pick a different key for the style.

If you chose local transcription or polish models, they download before the first run completes — that one-time wait is the model arriving, not the app's normal speed. Dictations after that run at full speed.

Yes — add the important names, recurring phrases, or replacements in Preferences and Acousmos will preserve them. The dictionary also learns from the corrections you make, so terms you fix stop coming out wrong.

Yes. Open Preferences and choose Run Setup Again — it walks through permissions, engines, and hotkeys from the start. The Setup Check panel in Preferences also shows the state of Microphone, Accessibility, the engines, and your hotkeys at a glance.