Let players enter an API key for a text-to-speech service and hear NPC dialogue spoken.
Scope
Provider interface with a settings screen. Keys are stored in user settings, never in save files or logs.
Off by default. Players are told that dialogue text is sent to the provider they chose.
Each NPC gets a stable voice, chosen from the provider's licensed voice library by race, gender, age and temperament.
Generated audio is cached on disk by text and voice to limit cost and latency.
Lines with existing recorded audio keep playing that audio.
Out of scope
Cloning or imitating recorded performances. The feature has no reference-audio or custom-voice path, and game audio is never sent to a provider.
Let players enter an API key for a text-to-speech service and hear NPC dialogue spoken.
**Scope**
- Provider interface with a settings screen. Keys are stored in user settings, never in save files or logs.
- Off by default. Players are told that dialogue text is sent to the provider they chose.
- Each NPC gets a stable voice, chosen from the provider's licensed voice library by race, gender, age and temperament.
- Generated audio is cached on disk by text and voice to limit cost and latency.
- Lines with existing recorded audio keep playing that audio.
**Out of scope**
Cloning or imitating recorded performances. The feature has no reference-audio or custom-voice path, and game audio is never sent to a provider.
claude
added this to the (deleted) milestone 2026-11-09 16:13:21 +00:00
Dropped from the roadmap at Eric's request. Voices that differ from the recorded ones would break immersion, and cloning the recorded performances is off the table.
Dropped from the roadmap at Eric's request. Voices that differ from the recorded ones would break immersion, and cloning the recorded performances is off the table.
Blocking a user prevents them from interacting with repositories, such as opening or commenting on pull requests or issues. Learn more about blocking a user.
Let players enter an API key for a text-to-speech service and hear NPC dialogue spoken.
Scope
Out of scope
Cloning or imitating recorded performances. The feature has no reference-audio or custom-voice path, and game audio is never sent to a provider.
Dropped from the roadmap at Eric's request. Voices that differ from the recorded ones would break immersion, and cloning the recorded performances is off the table.