zooqueen
2457efd855
The workspace already answers /chat and /dev. What was missing is the way in from everywhere else: a rail carrying the two modes, and the field that starts one, on all 945 pages. Both step aside inside the workspace, which draws its own sidebar with the same two modes and brings its own field — one table in components/shell/mode.ts says where that is, so a third mode would be one edit. The front page says one thing. It carried eleven sections and named the same claim in seven headings; it opens on the sentence now and answers the one objection that follows it. The description ~135 pages inherit listed eight nouns. Our own type says `site` where it said `surface` — NEXT_PUBLIC_SURFACE keeps its name, being the deploy lane's word rather than ours. Copy across 61 pages drops the feature lists that stood in for a position, and two live pages stop reaching for `powerful`.
4.1 KiB
title, slug, kind, status, version, updated, summary, order
| title | slug | kind | status | version | updated | summary | order |
|---|---|---|---|---|---|---|---|
| Biometric, Voice, and Likeness Notice | biometric-voice | notice | draft | 2026-07-22 | 2026-07-22 | How Hanzo handles faces, voices, and likeness in features that process them — notice, consent, retention and destruction, and the prohibition on unauthorized cloning. | 10 |
This Notice explains how Hanzo AI, Inc. handles biometric identifiers, voiceprints, facial data, and likeness in features that process them — for example voice synthesis and dubbing, avatar generation, lip-sync, and any identity or liveness verification. It supplements the Privacy Policy and, where applicable, satisfies notice requirements under laws such as the Illinois Biometric Information Privacy Act (BIPA), the Texas Capture or Use of Biometric Identifier Act, and Washington law.
1. When this Notice applies
It applies only to features that actually process a face, voice, biometric identifier, or likeness. Most Hanzo Services do not. Where a feature does, the flow presents this Notice and obtains the required consent before capture.
2. Consent before capture
We do not capture, generate, or use a biometric identifier, voiceprint, facial geometry, or a synthetic clone of a real person's voice or likeness without written consent obtained before capture, stated purpose, and the disclosures the law requires. You must not submit another person's face, voice, biometric identifier, or likeness for identification, cloning, dubbing, avatar creation, impersonation, or synthetic-media generation unless you have obtained that person's legally required permission and can produce it on request. This obligation is part of the Acceptable Use Policy.
3. Purpose limitation
Biometric and likeness data is used only for the specific feature you request — synthesis, dubbing, avatar rendering, lip-sync, or verification — and for the security and integrity of that feature. It is not used to identify you across contexts, sold or shared, used to infer characteristics, or used to train generalized Hanzo models absent the separate explicit permission described in the AI Research and Data Contribution Terms and the additional lawful condition that special-category data requires.
4. Retention and destruction
Biometric identifiers and templates are retained no longer than necessary for the feature and, in all events, no longer than applicable biometric law permits — under BIPA, the earlier of the purpose being satisfied or three years after the last interaction. We maintain a written retention-and-destruction schedule and destroy the data on that schedule. Generated media you create is retained under your account's ordinary retention settings; the underlying biometric template is destroyed per this schedule.
5. Synthetic media, provenance, and disclosure
Media that clones or manipulates a real person's voice or likeness is easy to misuse. Where you distribute it, you must disclose its synthetic nature as the law requires and preserve any machine-readable provenance or watermark the Services attach. Using synthetic media to defraud, defame, harass, sexually exploit, or interfere with elections is prohibited.
6. Disclosure to processors
Where a feature relies on a specialized vendor (for example, a liveness or identity-verification provider), that vendor is a subprocessor listed in the Subprocessor Register and is bound to equivalent obligations, including no independent training on biometric data.
7. Your rights
You may withdraw consent, request deletion, and exercise the other rights described in the Privacy Policy by contacting [email protected]. Withdrawal stops future processing and triggers destruction of the associated biometric data, subject to legal holds.
Publication gate: Do not enable a biometric, voice-cloning, or likeness feature under this Notice until the consent flow, the written retention-and-destruction schedule, the vendor DPAs, and counsel review (BIPA, Texas CUBI, Washington, and state biometric consent) are in place.