Generic scene
She entered the room and felt nervous. The silence was heavy.
Voice moat
Voice Studio turns style from a vague prompt into a reusable writing profile.
Upload samples, rebuild the profile, hear a scene in generic AI and in your voice, refine the result conversationally, then attach that voice to a project.
2
Hear It calls
Baseline and profiled voice comparison.
3
Public models
Claude Opus 4.7, Claude Sonnet 4.6, and Claude Haiku 4.5
BYOK
Cost clarity
Bring your own OpenRouter key; FictionOS does not mark up model usage.
Author voice
She paused at the threshold because the room had gone politely, professionally dead.
Character-filtered voice
Mara did not enter rooms. She negotiated with them until they admitted her.
It learns what you do and what you refuse to do.
A profile is more than adjectives. It records rhythm, prohibitions, exemplar passages, voice philosophy, and technique rules that can be reused across projects.
Hear It makes the moat visible.
The public value is the A/B moment: same scene, same premise, generic AI versus the profiled voice. That is much easier to trust than a paragraph saying "voice-matched."
Characterological filtering prevents voice imperialism.
The product has a character-voice layer that can arbitrate author voice against character cognition, so every character does not collapse into the narrator.
Privacy and cost stay explicit.
Voice work uses the connected OpenRouter key. Public copy should explain upload, rebuild, test, refine, attach, delete, and cost surfaces without promising legal outcomes.
Voice Studio is one of the easiest public proofs to understand: show the same scene before and after profile application, then let the writer decide whether the system hears them.