REFERENCE ARCHITECTURE
Architectural specifications for the CIP framework. These documents define the technical structures, API contracts, and operational models that the standard relies on.
Declaration Format Specifications
llms.md Format Specification
The lightweight flat-key format for declaring creative rights to LLM crawlers and retrieval agents.
ai.md Format Specification
The structured ten-section format for comprehensive, format-neutral AI rights declarations.
Architecture Documents
cip.md Field Specification
The formal field specification for the cip.md rights declaration format. The citable, versioned specification document consolidating all field definitions across v3.2–v3.29.
Lexicon Architecture
The framework’s lexicon is structured into three layers that distinguish what stability commitment applies to each entry, who maintains it, and how the broader community can contribute.
Mixed-Rights Framework
Addresses the structural limitation that both Cloudflare Content-Signals and the current cip.md format share: both treat content as having a single rights position per URL, with no mechanism to express different positions for different content layers.
NILP Architecture
Addresses NILP as the framework’s most cross-referenced concept, consolidating positions on NILP into a single canonical specification page covering the four primary rights, eight cip.md fields, and the NILP Downstream Obligation.
GIPL Insurance
Specifies the GIPL policy structure for the new exposure class created by AI content pipelines engaging subsisting rights at scale. Traditional liability policies do not match this exposure cleanly.
AI-IP Risk Scoring
The actuarial foundation that GIPL insurance pricing runs on. The metric the Underwriter Certification track is calibrated against and the scalar summary that portfolio reporting and premium-factor calculation reduce to.
RSE
Explaining the score in plain language. The companion architecture to the AI-IP risk index that provides structured explanations of risk scores for regulatory submission, board reporting, and policyholder communication.
CR-AII
Risk at portfolio scale. The portfolio-level companion to the AI-IP risk index that aggregates individual operator risk scores into sector-level and portfolio-level indices for insurers, reinsurers, and regulators.
Rights Registry
The operational verification layer that turns cip.md declarations from self-attested rights positions into third-party-verifiable rights records. Forward-looking specification; operational deployment pending governance decisions.
CDR Portal
The operator-facing web interface through which creators, agents, Authorised CDR Registrars, and Practitioners interact with the Rights Registry. Forward-looking specification.
Output-Provenance
When the result is generated, not fetched. The architecture for declaring rights provenance on AI-generated outputs — text synthesis, multimedia generation, federated/aggregated outputs — where the unit is produced at query time and layers are synthesised rather than authored.
Terms Enforcement API
The programmatic interface through which AI platforms query the framework’s declared rights positions in real time. Distinct from the Registry’s lookup API: resolves what a platform may do with a given asset.
Legal Annexes
The case law the framework anticipates. Nine annexes consolidating the major AI-and-IP litigation and policy developments that the framework’s rights architecture is designed to address.
Platform Implementation
Seven-step technical guide for platforms seeking CIP Platform Certification — from Rights Registry integration to annual audit readiness.
Standard Policy Wording
What the framework will procure. A workstream specification documenting what a Generative IP Loss (GIPL) standard policy wording must address, structured as a procurement contract for independent counsel rather than the wording itself.
Wiki Workstream Spec
What the framework will commission. A procurement specification for the collaborative glossary wiki infrastructure, documenting functional requirements, data model, editorial workflow, and technical stack rather than building the wiki speculatively.
Safeguarding Overlay
Architecture for protecting vulnerable individuals — minors, at-risk adults, and persons under witness protection or court-ordered anonymity — whose identity information must be suppressed from AI pipeline outputs even where rights otherwise subsist.
Subrogation Framework
After GIPL pays a claim, who bears ultimate responsibility? The framework architecture for upstream recovery pathways from AI model vendors, training data suppliers, platform operators, and infrastructure providers.
Auth Architecture
Forward-looking architecture for authentication and identity across framework services — the Rights Registry, CDR Portal, Terms Enforcement API, and Generator. Specification published; deployment pending Decision 1 (build vs integrate) and Decision 2 (identity provider selection) governance outcomes.
AI Underwriting Glossary
Sub-glossary of 30 terms with underwriting-specific structural data — category, module codes, definitions, and signal mappings — for the AI insurance questionnaire and AI-IP risk scoring engine. Operationally distinct from the main glossary because each term carries Category, Module letter codes, and Signal taxonomy significant for underwriters.
Sector Verticals & Topics
Film & TV Vertical
Architectural anchor for film and TV production rights work. Identifies seven workstreams covering substantive gaps in cip.md’s film/TV coverage. Workstream 1 (guild obligations) shipped at v3.32 with the CIP-Guild-Agreement field; workstreams 2–7 queued for subsequent releases.
Source Material Architecture
Architectural specification for declaring third-party copyrighted material present in the operator’s content. Nine category workstreams named and queued; CIP-Source-Material shipped as the v3.34 starter increment.
Advertising Vertical
Architectural specification for advertising vertical work. Three field families shipped (CIP-Ad-Disclosure, CIP-Campaign, CIP-Operator-Hierarchy); two further workstreams queued (targeting-data rights, syndication rights); RSE elevated to dedicated architectural treatment.
Video Vertical
Architectural specification for platform-native video content. Operationally distinct from Film & TV: addresses platform-mediated rights matching, monetisation, creator-vs-platform rights cascade, and platform-specific disclosure regimes. One field shipped (CIP-Video-Platform); six workstreams queued.
Photography Vertical
Architectural anchor for photography. Maps image copyright, subject likeness, AI training on archives, and capture-vs-synthetic provenance onto existing framework fields. Four workstreams queued for design.
Gaming Vertical
Architectural anchor for gaming. Maps multi-component works, performer voice/mocap consent, user-generated content, and AI-generated assets onto existing framework fields. Four workstreams queued for design.
Fashion Vertical
Architectural anchor for fashion. Maps design rights, marks/trade dress, model likeness, and AI-designed or virtual garments onto existing framework fields. Four workstreams queued for design.
Journalism Vertical
Architectural anchor for journalism. Maps TDM/training ingestion, summarisation vs substitution, attribution/byline, source protection, and syndication onto existing framework fields. Four workstreams queued for design.
Brand Vertical
Architectural anchor for brands. Maps trademarks/trade dress, AI style/voice mimicry, impersonation/passing-off, and brand asset use onto existing framework fields. Four workstreams queued for design.
Indigenous IP
Provisional problem-framing only — not a standard. Conventional IP maps poorly onto Indigenous knowledge and traditional cultural expressions. Any framework treatment must be developed with Indigenous communities, not authored for them, under free, prior and informed consent.