[Task]: Improve Edge TTS voice selection during setup

[crankshift]

Summary

Improve the setup command flow for Edge TTS voice selection so users can keep the default voice or choose a different voice through guided, validated selection.

Context

When a user selects Edge TTS during setup, the current flow should offer a default voice path and an alternate voice selection path. If the user chooses another voice, setup should first present a country list with autocomplete. Free text must not be submittable; users should only be able to continue after selecting a country from the list. After country selection, setup should show a paginated voice list for that country because Edge TTS enforces pagination for countries with large voice lists. The voice selection should support keyboard navigation and autocomplete.

Acceptance Criteria

• Edge TTS setup prompts the user to keep the default voice or select another voice.
• Selecting another voice opens a country selector with autocomplete and keyboard navigation.
• Country selection rejects arbitrary free text and only accepts an item selected from the provided list.
• After country selection, setup shows a paginated list of available voices for that country.
• Voice selection supports keyboard navigation and autocomplete.
• Pagination handles countries with large voice lists without attempting to fetch or display all voices at once.
• The selected voice is saved in setup configuration.
• Tests cover default selection, country validation, paginated voice selection, and saved config output.

Notes

Edge TTS API pagination is a constraint for countries with large voice lists. Keep this flow inside setup/config UX and avoid hardcoding a static global voice list if the API is the source of truth.