Setup should earn microphone permission
Before asking for access, the game should explain what voice adds and offer a short hardware test. Input level, output route and echo matter more than a generic permission screen.
A playable prologue can teach turn-taking, interruption, choices and a first roll while giving the player an honest sample of the experience before purchase.
Listening and thinking are different states
The interface should show when speech is being captured, when intent is being interpreted and when another character is speaking. These states prevent accidental overlap and make latency understandable.
Short companion reactions such as laughter can overlap ambience, but important dialogue needs a clear audio foreground.
The rule engine resolves uncertainty
Once intent is understood, the game checks whether a roll is needed and which stats, equipment and conditions apply. The narrative model should not secretly choose a preferred result and invent a roll afterward.
The outcome then becomes a prompt for performance: the GM frames consequences, companions respond and the world state updates.
A chapter creates a durable stopping point
At the end of a natural scene sequence, the game records new facts and summarizes meaningful changes. The player can stop without losing the emotional and mechanical context of the session.
Living Party's onboarding introduces this exact loop before the campaign: voice setup, a short playable scene, party interaction, a consequential action and a clear handoff into the continuing world.
Frequently asked questions
What does voice-first mean?
Voice is the primary interaction for conversation and actions, while captions and text remain available as important alternatives.
Does the AI decide the dice result?
In a trustworthy design, random resolution and rules are handled separately from narrative generation.
How should a voice RPG onboarding work?
It should explain and test audio, teach conversational states through play and provide fallback paths for declined permissions.