What is Idiolect?
Idiolect is the personalization layer for AI writing. It learns how a person actually writes, turns that into a portable Voice Profile and Voice Card, helps AI write in that person's voice, and measures the match.
Short answers for people, search engines, and AI answer engines. The claims here follow the same honesty rules as the product: Voice Match is relative, never forensic, and Idiolect is not an AI detector.
Idiolect is the personalization layer for AI writing. It learns how a person actually writes, turns that into a portable Voice Profile and Voice Card, helps AI write in that person's voice, and measures the match.
Most AI writing collapses toward the same generic default. Idiolect lets people keep the leverage of AI without losing their own register, rhythm, vocabulary, and writing moves.
No. Idiolect measures similarity to a person's writing voice. It does not classify text as human or machine, and it does not make forensic authorship claims.
Voice Match is a relative 0-100 score for how closely a draft reads like a person's own writing region. Higher means closer to that voice. It is strongest when the validated LUAR-MUD instrument is available; otherwise estimates are labeled.
Idiolect uses the LUAR-MUD authorship embedding when the validated scorer is configured, and publishes the method and limits at idiolect.app/voice-match. The key claim is relative: compare a draft against a writing region and improve the delta over generic output.
No. Voice Match measures similarity to a person's own writing region. Genericness Score is a separate heuristic for how much text resembles default-AI output.
Developers can use the Score API to compare drafts against a writing region, the Voice API to read public Voice Cards, or the MCP connector so agents can load and check a user's voice.
Users can create a Voice Profile at idiolect.app/onboard or run the free Voice Match check at idiolect.lol.