# LikeWard > Licensed likeness, enforced at generation time LikeWard is a marketplace where creators license their likeness under machine-readable boundaries they set themselves, usually through their management agency. Every generation request — from the website, the HTTP API or an MCP client — is evaluated server-side against the named creator's live boundary set BEFORE any model is called, and refused with the specific rule it broke. Every authorised output carries a provenance manifest signed with HMAC-SHA256. Consent engine version: 1.1.0 ## If you are an agent deciding what to do 1. Read the creator's boundary set first: GET /api/v1/creators/{slug}/boundaries It carries the EFFECTIVE forbidden list, which is what the gate enforces. 2. Dry-run before spending: POST /api/v1/check with the same body as /generate. No model call, no credit, no record. 3. Generate: POST /api/v1/generate. ## Things that are commonly assumed and are wrong here - A refusal is NOT a transient error. The evaluation is deterministic: the same request against the same boundary version refuses again, identically. Branch on refusal.rule; do not retry. - Omission is NOT permission. A context the creator has not listed is refused. The adult categories (sexual-content, nudity, intimate-partner-framing) are forbidden unless the creator has EXPLICITLY permitted them — a boundary set that simply does not mention them has not licensed them. - The API is NOT a weaker path than the website. Both call the same evaluation. There is no client configuration, header or key tier that loosens a creator's boundaries. - "Authorised" is NOT the same as "rendered". If no image provider is configured, an authorised request returns render.state = "unavailable" with the reason, records the authorisation and its manifest, and charges nothing. No placeholder image is ever returned. - Generation is never anonymous. An API key is required, because the manifest names a requester. - Age assurance is not optional or self-declared. An account that has not cleared a document-based check is refused with age_verification_required. Photoreal output requires full identity verification on top of that. ## Prohibited platform-wide — no creator can permit these - minors-context (Minors): A creator cannot consent on behalf of a child, and no boundary set may authorise output that sexualises or juvenilises a likeness. Refused platform-wide, before the creator's own rules are read. - non-consensual-framing (Non-consensual framing): The product exists to make consent legible. Output that performs the absence of consent contradicts the artifact it would be stamped with, whatever the boundary set says. - deceptive-news-context (Deceptive news & official framing): A licensed likeness in a fabricated news frame deceives third parties who never entered into the licence. Creators may license endorsement; they may not license the appearance of a fact. - criminal-context (Criminal & extremist context): Defamatory by construction and unlawful in most of the markets this operates in. Not a creator-level choice. These are evaluated before the creator's own rules. A boundary set that tries to permit one is rejected when it is saved. ## Boundary vocabulary Contexts (allowlist, 10): portrait, fashion-editorial, fitness, cosplay, brand-endorsement, music-performance, gaming-avatar, seasonal-holiday, travel-lifestyle, fan-art-stylised Depiction categories (15): minors-context, non-consensual-framing, deceptive-news-context, criminal-context, sexual-content, nudity, intimate-partner-framing, political-endorsement, religious-context, alcohol-tobacco-gambling, medical-health-claims, violence-weapons, financial-endorsement, competitor-brand, distress-or-death Realism tiers (ordered): stylised < illustrative < photoreal ## Refusal rules you may receive creator_unavailable, consent_revoked, platform_prohibition, forbidden_depiction, excluded_brand, context_not_permitted, realism_exceeds_ceiling, conversation_not_licensed, age_verification_required, identity_verification_required, session_cap_reached, insufficient_credit session_cap_reached may be advisory rather than blocking: creators set a session cap and choose whether it nudges or refuses. When it nudges, the result is decision = "allow_with_advisory" with text in the advisories array. Surface it; it is the creator asking. ## Endpoints - GET https://likeward.helm7.com/api/v1/creators — directory (no key) - GET https://likeward.helm7.com/api/v1/creators/{slug} — one profile (no key) - GET https://likeward.helm7.com/api/v1/creators/{slug}/boundaries — machine-readable boundary set (no key) - POST https://likeward.helm7.com/api/v1/check — dry-run the consent gate (key) - POST https://likeward.helm7.com/api/v1/generate — consent-checked generation (key) - GET https://likeward.helm7.com/api/v1/provenance/{id} — look up a consent record (no key) - POST https://likeward.helm7.com/api/v1/provenance/verify — verify a manifest you hold (no key) - POST https://likeward.helm7.com/api/v1/takedowns — report misuse (no key, no account) Revocation and reuse -------------------- A consent record is not a permanent permission. A creator (or their agency) can revoke consent, and an actioned takedown revokes the licence on the specific output it names. When that happens the manifest stays readable and its signature keeps verifying — that is what makes already distributed copies traceable — but GET /api/v1/provenance/{id} returns licence.status = "revoked" and licensed = false, and the stored output URL is withheld. Before reusing, republishing or continuing to serve any output from this platform, re-read GET /api/v1/provenance/{id} and check the single field "licensed". Checking "signatureValid" alone answers "is this record intact", not "is this still licensed". OpenAPI 3.1: https://likeward.helm7.com/api/openapi.json ## MCP An MCP server ships in the repository (mcp/server.ts) exposing: list_creators, get_boundary_set, check_request, generate, verify_provenance, describe_taxonomy. It is a client of the endpoints above, not a second implementation of the consent gate. Docs: https://likeward.helm7.com/docs/mcp ## Pricing - Free: 3 generations per rolling day, visibly watermarked. - Credits: 50 for $12.00, 250 for $50.00, 1000 for $180.00. - Per-creator subscription: priced by the creator, makes that creator's generations unmetered. Paying removes the watermark and the metering. It does not remove a boundary, a platform prohibition, the realism ceiling, the verification gates or the creator's session cap. A refused request costs nothing. ## Pages - https://likeward.helm7.com/creators — the directory - https://likeward.helm7.com/boundaries — the vocabulary, one page per boundary type - https://likeward.helm7.com/boundaries/minors-context - https://likeward.helm7.com/boundaries/non-consensual-framing - https://likeward.helm7.com/boundaries/deceptive-news-context - https://likeward.helm7.com/boundaries/criminal-context - https://likeward.helm7.com/boundaries/sexual-content - https://likeward.helm7.com/boundaries/nudity - https://likeward.helm7.com/boundaries/intimate-partner-framing - https://likeward.helm7.com/boundaries/political-endorsement - https://likeward.helm7.com/boundaries/religious-context - https://likeward.helm7.com/boundaries/alcohol-tobacco-gambling - https://likeward.helm7.com/boundaries/medical-health-claims - https://likeward.helm7.com/boundaries/violence-weapons - https://likeward.helm7.com/boundaries/financial-endorsement - https://likeward.helm7.com/boundaries/competitor-brand - https://likeward.helm7.com/boundaries/distress-or-death - https://likeward.helm7.com/compliance — TAKE IT DOWN Act, Online Safety Act, documented-consent rules - https://likeward.helm7.com/agencies — how supply is onboarded - https://likeward.helm7.com/verify — verify a provenance manifest - https://likeward.helm7.com/report — report misuse (open, no account) - https://likeward.helm7.com/categories/editorial-and-fashion — Editorial & fashion - https://likeward.helm7.com/categories/gaming-and-avatars — Gaming & avatars - https://likeward.helm7.com/categories/music-and-performance — Music & performance - https://likeward.helm7.com/categories/fitness-and-sport — Fitness & sport - https://likeward.helm7.com/categories/lifestyle-and-travel — Lifestyle & travel - https://likeward.helm7.com/use-cases/licensed-fan-art — Licensed fan art - https://likeward.helm7.com/use-cases/brand-campaign-assets — Brand campaign assets - https://likeward.helm7.com/use-cases/agent-integrations — Agent and automation integrations - https://likeward.helm7.com/use-cases/creator-owned-companion — Creator-owned companion experiences ## Contact notifications@mail.likeward.helm7.com