Three meanings of voice control
Command control listens for a narrow action such as 'roll the dice'. Spoken-choice control accepts one of the paths the author has offered, often with synonyms. Free-form participation lets you phrase an action more naturally and asks the system to interpret it. PlayNook documents the first two; TWIST's guide explains predefined player utterances; Playfl presents the third as natural participation inside a scene.[1][2][3]
| Mode | Player freedom | Main design risk |
|---|---|---|
| Fixed command | Low | You must remember supported words |
| Spoken authored choice | Medium | Recognition can hide the available route |
| Free-form reply | High | The story may misread intent or lose structure |
Recognition is not the same as generation
A game can recognise speech without generating any story text. TWIST describes a tree of authored segments connected by declared utterances and synonyms. Conversely, a system can generate a story from typed input without using speech recognition. Input, interpretation and story output are separate layers.[2]
Gydel makes another useful distinction: a spoken action is added as a ready-to-run choice, then the player confirms or replaces it. That extra confirmation can reduce accidental actions while preserving broader phrasing.[4]
What to check before granting microphone access
Browsers require permission before a site can capture microphone media, and getUserMedia is limited to secure contexts. A trustworthy start flow should explain why it wants the microphone, provide a clear alternative where possible, and avoid implying that access is required when a text mode can perform the same job.[5]
- Does the story explain what will be captured before the browser prompt appears?
- Can you revoke access or finish the story by text?
- Does the product say whether speech is matched locally, sent to a service or stored?
InnerPlay's public build does not use voice input
InnerPlay's intended format is conversational and audio-first, but this production release is a text-and-caption prototype. You type the answer. The character's replies are placeholder text. Nothing here should be read as a claim that the current website offers live voice conversation.
Sources checked 19 August 2026.
Questions, answered
- What is a voice-controlled story game?
- It is an interactive story in which speech is an input. The game may listen for fixed commands, authored choices or freer natural-language actions.
- Does voice control mean the story is AI-generated?
- No. Speech recognition can route to fully authored branches, while a generative story can accept text instead of speech.
- Does InnerPlay have live voice today?
- No. The current public build uses typed answers and captioned placeholder replies.
Sources
- PlayNook FAQ — PlayNook. Checked 19 August 2026.
- How to Write Interactive Fiction and Audio Stories — TWIST Tales. Checked 19 August 2026.
- Playfl — Playfl. Checked 19 August 2026.
- Gydel — Gydel. Checked 19 August 2026.
- MediaDevices: getUserMedia() — MDN Web Docs. Checked 19 August 2026.