Find practical answers before you install, compare, or choose a plan.
End-to-end setup for Ubuntu, Fedora and Arch. Covers Wayland vs X11, pipewire, input methods, and the first dictation loop.
# 1. Confirm audio stack $ pactl info | grep -i 'Server Name' # 2. Detect session $ echo $XDG_SESSION_TYPE # 3. Run compatibility report $ aispeaky doctor
Cards are grouped from the most-requested topics. Filter above to narrow.
Which surfaces expose insertion APIs today, where command mode is gated, and how to tell which session you are on.
Where the desktop apps stand, what the waitlist actually unlocks, and how managed lanes will differ from the Linux build.
How the on-device baseline differs from opt-in managed accuracy, what stays on your machine, and when to switch lanes.
Plain-language list of certifications, residency claims and platform guarantees that are not in scope for v1.
Where each tool fits today: dictation accuracy, supported platforms, command coverage, and licensing model.
Setup time, model management, latency, and the boring maintenance bits a polished desktop client takes off your plate.
Where voice helps today, where it does not, and how to combine dictation with keyboard breaks without medical overclaims.
The two permission paths on Linux, the most common pipewire conflict, and how to confirm the right input is selected.
Why insertion can fail per app, how to confirm focus, and which input-method combos are known to break.