How it works

Avatars are how people show up online now.

Most online interaction happens through a represented self — a name and picture, a curated profile, a creator persona, a voice, and increasingly an AI stand-in that can act on someone's behalf. RSP treats that avatar as something the person owns and controls, not something a platform can quietly reuse. See the full set of tracked consent dimensions on the Dimensions page.

Select a step to see how it works ↓

01

Claim — binding your represented self

Claiming establishes that a given avatar is yours. Your display name, picture, likeness, recorded voice and any AI stand-in are linked to your core identity in the RSP model, so every later grant references one owner. Nothing is shared at this stage — claiming is about ownership, not exposure. This is the identity-context layer described under Consent modules, applied to representation rather than raw account data.

02

Grant — scoped, revocable permission

A grant lets a specific context project part of your avatar. Each grant is scoped by where it applies (which hub or product), how it may be used (view-only likeness, live voice, an acting AI twin), and for how long (a single session, a fixed window, or until revoked). Grants are additive and never permanent — likeness and voice are always time- or session-bounded rather than a blanket hand-over.

03

Project — rendering only what you allowed

When others encounter you, RSP renders only the facets your active grants permit. Presence signals control how your avatar appears live — available, busy, quiet — so being seen never means being surveilled. Identity-exposure protection guards against unwanted visual or voice similarity, keeping an AI representation clearly tied to, and bounded by, your consent.

04

Revoke — withdrawing a represented self

Revocation withdraws a grant the same way it withdraws any other RSP permission. The projection stops everywhere the grant reached, AI stand-ins lose the authority to act or speak for you, and the change is recorded in the consent ledger without becoming a surveillance dashboard. Consent is a living state you can end at any time, not a one-time checkbox.

Worked example

Maya is a recovery coach who records short encouragement clips. Here is how her represented self moves through RSP — from claiming her avatar to revoking it.

  1. 1
    Claim

    Maya links her name, photo, recorded voice and a permitted AI voice model to her core identity. Nothing is shared yet — she simply owns her represented self.

  2. 2
    Grant

    She grants the Recovery Circle hub a session-scoped voice grant so her AI stand-in can read new encouragement messages aloud during evening check-ins only.

  3. 3
    Project

    During check-in, members hear Maya's voice on today's message and see her presence as available. They cannot access her likeness elsewhere or generate new clips — RSP renders only what she allowed.

  4. 4
    Revoke

    When Maya pauses coaching, she revokes the grant. The AI voice can no longer speak for her, past projections stop, and the change is logged in her consent ledger.