F. Elusio Zafar & Jeff Gordon (Founders of Creation Rights)
team@creationrights.ai
www.creationrights.ai
PDF download: https://drive.google.com/file/d/1lTrOXJLvtMTRHw3avZ-BIsm-56CMEt7X/view?usp=sharing
Abstract. The Creation Rights Protocol, is a metadata-native rights system designed for the modern human and AI content ecosystem. Grounded in dynamic metadata architecture, this protocol enables creators to attach, update, and enforce rights in real time using AI-readable metadata, NIL attribution for publicity rights, and decentralized registries.
Creative consumption and work is shifting the internet from a static publishing model to a dynamic, redistributable ecosystem fueled by AI and user-generated content by modifying, remixing, or personalizing the content being consumed. However, content metadata systems remain outdated, brittle, and ineffective at enforcing creator rights, especially in environments where real-time transformation and distribution dominate.
Unlike traditional copyright, which relies on passive text and reactive legal measures, the Creation Rights protocol embeds active, dynamic enforcement logic in content itself; transforming the file into an AI-compliant, licensable, and trackable digital asset. This system is accessible over common internet protocol, creating a neutral metadata fabric for the entire internet and paving the way for a fairer, creator-first internet.
Today’s content is remixable, AI-trainable, and instantly distributable, however, the rights infrastructure remains outdated, still relying on static file headers, inaccessible legal PDFs, and unenforceable screenshots. Metadata today is static, often generated once and rarely updated. It is brittle, meaning it can be easily removed, ignored, or misread. It is non-enforceable, as it lacks a direct connection to licensing or usage rights. And it is platform-specific, typically buried within siloed systems. As a result, content metadata is often ignored or stripped by major platforms. It is incompatible across file formats, blind to how content evolves or is remixed, and ineffective for AI systems that operate at scale.
At the same time, content itself has evolved. It is now multimodal, including voice, video, 3D, images, and biometric data. It is increasingly generated and remixed by humans, AI models, and hybrid systems. And it is monetized across diverse platforms, from social media to commercial remix ecosystems. To keep up, we need metadata that lives and breathes with the content it describes and protects.
Creation Rights solves this problem by providing dynamic, structured metadata that reflects the current licensing status. It supports real-time updates to a single source of truth based on actual content use. It includes embedded licensing logic that is readable by APIs, AI systems, and digital marketplaces. And it ensures metadata is cryptographically verifiable through linked hashes and watermarks.
The Creation Rights protocol architecture is built around a modular, metadata-first stack that integrates with modern cloud storage, machine-readable metadata formats, AI inference systems, and legal rights infrastructure. At its core is a dynamic metadata engine that ingests creative media, extracts descriptive and structural data using AI tagging, assigns licensing attributes, and embeds enforceable metadata across the file and network layers.
The system begins with secure OAuth-based storage integration, allowing users to connect file storage such as Google Drive, or optionally IPFS for decentralized storage. Once a file is uploaded, the file detection engine identifies its MIME type, format, and media classification (e.g., image, video, audio, text, or 3D asset). This enables context-aware processing by downstream modules.
The NIL attribution layer assigns Name, Image, and Likeness profiles, which can include biometric traits such as voiceprints, visual avatars, and creative signatures. This feature is essential for agencies or creators representing multiple identities or assets across different works. The AI tagging module then analyzes the asset’s underlying structural components, format, genre, and stylistic indicators using custom-trained models and large language understanding, resulting in a rich metadata profile that includes genre tags, emotional tones, and usage contexts.
The metadata engine generates both human-readable and machine-readable metadata in standard formats like agent.json and cr.txt. These files conform to open web schemas (JSON-LD, RDFa, and IPTC/XMP/ID3 depending on the media type) and include fields for creator identity, license type, rights terms, and remix lineage. Metadata is either embedded directly into the media file (where supported) or hosted externally and linked via canonical URLs.
The Creation Rights protocol provides licensing options from default (All Rights Reserved) to flexible commercial terms like Creative Commons and Exclusive Licenses. Each asset is associated with a royalty structure and programmable logic to track downstream use. Once licensed, the audit and delivery layer logs all access events, processes payment flows, and provides the buyer with a secure, watermarked ZIP package that includes the master file, license, and metadata.
Each component is designed to be modular, interoperable, and verifiable, creating a complete rights infrastructure that integrates seamlessly with the AI era and creator economy.
The dynamic metadata workflow lies at the heart of the Creation Rights protocol, transforming raw creative assets into structured, rights-enforced digital objects that are both machine-readable and legally traceable. This end-to-end process begins when a creator uploads a file from an integrated cloud storage provider, such as Dropbox or Google Drive, or directly calling the Creation Rights API. Once the file is ingested, the system performs MIME-type detection and file classification to determine whether the asset is an image, video, audio file, document, or 3D model.
Following this, the system prompts the creator to assign or create a NIL (Name, Image, and Likeness) profile, which encapsulates key identity markers associated with the work; such as the creator’s voice, likeness, or motion traits. These NIL profiles may be reused across projects or segmented for teams and agencies managing multiple creative identities.
The protocol then applies AI-assisted tagging using a pipeline of computer vision e.g. executing object detection, semantic segmentation, and natural language models. The protocol then applies AI-assisted tagging using a pipeline of computer vision and natural language models. These models extract metadata such as genre, aesthetic style, mood, format, and theme, along with contextual signals useful for discovery and licensing. For example, natural language models apply sentiment analysis to accompanying text, captions, and transcripts to determine emotional tone (e.g., joy, tension, melancholy), assigning probabilistic confidence scores to each detected sentiment. Image and audio classifiers use embedding vectors and model-attention weights to score attributes such as vibrancy, aggression, harmony, or suspense, based on training datasets curated by creative industry taxonomies. The protocol captures these scores in the AI Routing Layer of the metadata schema, offering structured outputs like a "Mood Intensity Index" or "Stylistic Diversity Score" that platforms and models can use to optimize recommendations, remix decisions, or rights-based filtering. Creators can review, refine, or augment these tags within the metadata interface.
The next step is the IP rights definition. The creator selects a licensing scheme from a default list, ranging from All Rights Reserved to Commercial, Exclusive, or Creative Commons (e.g., CC-BY-NC). Each license selection results in both a human-readable summary and a machine-readable ruleset embedded into the content’s metadata file. These rights terms also govern downstream usage, remix eligibility, and revenue-sharing logic.
Once the license and metadata are confirmed, the system generates standardized metadata packages (agent.json, cr.txt) that contain all asset descriptors, creator identifiers, version hashes, usage terms, and lineage links. This metadata is either embedded into the media file (e.g., via ID3 tags for audio, IPTC/XMP for image/video) or hosted externally and referenced via <link rel="license"> tags or API endpoints.
To protect the asset during distribution, a watermarked preview is automatically synced or generated using e.g. Sharp.js for images, e.g. SoX for audio, and e.g. ffmpeg for video. These previews are clipped, muted, or overlayed to prevent unauthorized reuse while still enabling promotion and search.
The asset is then published and made available for licensing. Every interaction, whether preview, license request, download, or remix, is logged and can be with a cryptographic audit trail and versioned metadata chain. If a file is updated, reused, or remixed, its metadata evolves accordingly, creating a dynamic and enforceable record of creative intent.
Data flow:
The metadata specification defines the core structure that Creation Rights uses to describe, license, and track digital content. Built for both human and machine consumption, this metadata system draws from open semantic standards like JSON-LD, RDF, and IPTC, while adding domain-specific fields required for AI interoperability, NIL attribution, and royalty enforcement. It is designed to be flexible enough to describe any form of media while remaining consistent across file types and licensing models.
Each piece of content that enters the Creation Rights pipeline is assigned a unique identifier, which can be expressed via a cryptographically verifiable fingerprint. This fingerprint anchors the Identity Layer, which includes creator information, NIL profile hashes, and unique asset identifiers. This identity layer enables creators to tie each file to a legal or biometric entity, and provides AI systems with a trusted source of origin.
The Structural Layer includes core technical descriptors of the content: MIME type, resolution, duration, aspect ratio, bitrate, and other codec-level metadata. These fields help platforms, search engines, and AI systems sort, render, and process content appropriately, while also helping remixers or collaborators understand technical constraints.
The Legal Layer encodes the licensing terms associated with the asset. This includes the selected license template (e.g., All Rights Reserved, CC-BY-NC, Exclusive), usage permissions, remix allowances, and commercial restrictions. The legal layer also includes licensing identifiers that link to agent.json or cr.txt endpoints hosted by Creation Rights, as well as optional smart contract references that can trigger automated enforcement, access control, or royalty distribution.
To support AI usage and discovery, the AI Routing Layer adds higher-order metadata including genre tags, inferred thematic context, stylistic dimensions, and remix depth. These tags are generated and auto-assigned by AI models trained on cultural, genre, and sentiment analysis and are mapped to controlled vocabularies that improve searchability and model alignment.
The Audit Layer is responsible for traceability and compliance. It includes a changelog of metadata versions, SHA-256 hashes of every previous state, and links to associated usage logs. This creates a tamper-proof lineage trail for every asset, allowing for dispute resolution, usage validation, and royalty audits.
Each of these layers can be embedded inside the content (e.g., EXIF, ID3 for audio, XMP/IPTC for visual media) or published as externally hosted metadata documents accessible via persistent URLs. This dual-mode architecture ensures that the metadata remains portable, verifiable, and enforceable across both legacy systems and future-facing platforms.
Royalty Engine, Routing and Payment Logic:
The Creation Rights protocol also includes a built-in Royalty Engine that enables automated revenue sharing among collaborators. Creators can assign percentage-based splits to different NILs or wallet addresses. These assignments are stored as part of the asset’s metadata and versioned to ensure accurate lineage over time. At the point of purchase, license issuance, or usage reporting, royalties are automatically routed through integrated financial services based on the latest rights snapshot. This ensures that every transaction is settled according to the creator-defined logic, whether single-pay, subscription, micro-royalty, or donation-based.
The Royalty Engine in the Creation Rights protocol is a programmable, metadata-driven payout system designed to facilitate transparent, real-time distribution of revenue among collaborators. It integrates deeply with the licensing and attribution infrastructure, ensuring that the logic of creator compensation is embedded in the same metadata layer that governs rights and usage.
When an asset is licensed, downloaded, or monetized, the protocol executes a royalty distribution plan encoded in the asset's metadata. This plan includes a list of royalty recipients, typically tied to NIL profiles or verified wallet addresses, and the associated percentage each party should receive. These rules are immutable for the given asset version and are validated on every transaction or usage event.
Under the hood, the Royalty Engine supports multiple payment configurations:
Payments are routed through secure, audited services, and decentralized routing partners. The system is extensible to accept crypto wallets and stablecoin payments where jurisdictionally permitted. Each transaction generates a metadata receipt, including license ID, date and time of transaction, royalty allocation breakdown, and payment processor record or hash.
These receipts are stored within the audit trail and can be queried by creators, buyers, auditors, or governing bodies. This allows for reconciliation, reporting, and dispute resolution with a level of transparency previously unavailable in traditional IP markets.
The Royalty Engine is designed to scale from solo creators to production teams, record labels, game studios, and distributed collectives, supporting both micro-transactions and enterprise-grade licensing flows.
License Selection and Automation:
Upon asset ingestion, creators are presented with license options that allow them to define how their work can be used. These include the pre-configured templates:
In addition to manual selection, the system uses AI tagging and usage prediction to suggest appropriate licenses based on content type, audience, or prior licensing behavior. For example, a creator known for releasing open datasets may be prompted to use CC0 by default, while a commercial photographer might be offered a tiered licensing model with editorial vs. advertising rates. Licensing logic is recorded in machine-readable format (agent.json) and is automatically embedded in metadata previews, zip packages, and delivery receipts.
Licensing and attribution are core to the Creation Rights protocol, providing creators with granular control over how their content is used, shared, and monetized across digital ecosystems. Rather than treating licensing as a separate document or post-hoc agreement, Creation Rights embeds licensing logic directly into the metadata of each file, making rights enforceable by both humans and machines.
At the point of content registration, creators are guided through an intuitive License Selector UI. This interface allows them to choose from a range of standardized license templates, including "All Rights Reserved," "Commercial," "Exclusive," and "Creative Commons Attribution-Non Commercial (CC-BY-NC)." For power users and legal teams, custom licenses can be authored or uploaded. Each license includes predefined fields for usage type, modification rights, commercial terms, territorial scope, and exclusivity duration.
These licensing parameters are then encoded into both human-readable summaries and machine-readable documents, specifically in the agent.json and cr.txt formats. The documents are structured to be interoperable with industry standards such as the Programmable IP License (PIL) and RDFa-compliant rights schemas. Each license entry also generates a unique License ID, which is included in all associated logs, watermarks, and downloadable ZIP bundles.
Complementing the licensing system is a robust NIL (Name, Image, and Likeness) attribution module, which enables creators to link their biometric and stylistic identity to assets. NIL profiles can include structural elements like voiceprints, facial features, motion signatures, and stylistic elements. These traits are cryptographically hashed and stored for example via IPFS, providing a verifiable, immutable record of identity. This system allows agencies, teams, or pseudonymous creators to manage multiple NILs under a single account or organizational identity.
Every licensed asset is published with a unique, persistent metadata URL in the form https://creationrights.com/agent/{creator}/{asset}. This metadata endpoint serves as a real-time license and attribution reference that can be queried by content platforms, search engines, and AI models to verify rights status, NIL ownership, and permitted usage.
Together, these systems ensure that licensing is not just a passive legal placeholder but an active, programmable and enforceable mechanism for protecting creators' work in a real-time, remixable internet:
The Creation Rights enforcement layer transforms passive metadata into active digital infrastructure for rights protection. Rather than depending solely on legal threats or post-violation takedowns, the protocol embeds real-time, programmable enforcement into the content lifecycle, making unauthorized use visibly deterred, traceable, and reversible. The enforcement layer is built pillars of visibility, auditability, and automation.
The protocol syncs or applies smart watermarking to every preview version of a protected asset. Images are overlaid, audio tracks are muted or watermarked with inaudible tones, and video is clipped and visually tagged. These previews serve as both promotional tools and protective wrappers, ensuring that valuable IP cannot be misused even before licensing. Importantly, these transformations are non-destructive and tied directly to the licensing state, ensuring that licensed users receive high-resolution, unmarked master files.
Compliance logging is implemented at every point of interaction with an asset. Each download, preview, remix, or delivery event is recorded with a timestamp, IP address, License ID, and metadata hash. These logs are stored in secure, append-only systems to ensure tamper resistance. This creates a cryptographic chain of custody for every asset, enabling forensic validation and traceability across usage events.
The protocol supports real-time validation via webhooks and API endpoints. Platforms, marketplaces, or AI systems can query a metadata URL to confirm license validity before rendering or ingesting a file. A content usage engine compares inbound requests to license terms in agent.json and can allow, restrict, or prompt negotiation for additional rights. This allows AI training systems, content platforms, and decentralized networks to filter their ingest based on up-to-date legal compliance.
The protocol includes an automated takedown and dispute system. Through the Creation Rights dashboard, creators can generate DMCA notices, submit on-chain complaints, or challenge unauthorized NIL usage. These actions can trigger platform-specific enforcement actions or escalate to governance modules for community adjudication. Trusted admin roles can override or edit metadata in cases of dispute resolution or verified misattribution.
Future roadmap features include integration with digital fingerprinting databases, remix trace detection, and DAO-based enforcement mechanisms. Combined, these capabilities offer creators a dynamic, programmable enforcement toolkit that scales across platforms and jurisdictions; without requiring gatekeepers or litigation. The protocol will allow for the future enforcement layer features, Smart Previews: Downsampled, clipped, muted, or overlaid; Compliance Logs: License ID + timestamped delivery receipt; Webhook Hooks: Allow AI models or platforms to verify rights; and Takedown UI: DMCA generator, blockchain ownership validation, and NIL reassignment.
Creation Rights allows creators to protect and license their media with automated metadata. It enables AI companies to train on legally cleared media with verified lineage. Marketplaces can sell digital goods that have embedded royalty logic. Developers can use the Creation Rights API to build compliance-aware applications.
For creators, Creation Rights offers a seamless method to secure, license, and monetize their work from the moment of creation. Whether uploading a song, short film, dataset, or 3D animation, artists can automatically generate enforceable metadata, attach licensing terms, and embed a royalty model that pays collaborators over time. This eliminates the need for separate watermarking services, manual copyright filings, or dependency on centralized marketplaces. Moreover, the NIL attribution layer allows creators to protect and manage their voice, likeness, or style, even in synthetic or AI-generated works.
For AI developers and companies, Creation Rights offers a machine-readable framework for sourcing ethically and legally licensable training data. AI systems can query the metadata endpoints of media files to confirm training permissions, verify attribution requirements, and respect opt-out or non-commercial clauses. This prevents inadvertent copyright violations and establishes a content provenance trail for every model training step. Generative AI systems can also publish outputs with embedded dynamic metadata, allowing downstream users to trace lineage and comply with source licenses.
For digital platforms and marketplaces, Creation Rights serves as an attribution-aware compliance layer. Streaming services, creator marketplaces, and remix platforms can query and respect embedded metadata to automatically enforce license limits, such as commercial vs. non-commercial use, remix eligibility, or embargo dates. This reduces legal risk and increases trust among users by transparently honoring rights at scale. For platforms with royalty-based payout models, the protocol’s integrated metadata and audit trail simplifies accounting and revenue distribution.
For developers, Creation Rights provides a RESTful API, webhook endpoints, and open SDKs to build custom integrations. This includes CMS plugins, creator dashboards, compliance filters, and remix-aware publishing tools. Developers can access the underlying metadata schemas to build analytics, search engines, or licensing layers optimized for rights-aware interaction.
For educators and open-source communities, Creation Rights makes it easier to publish content under specific educational or attribution licenses. Metadata can include citations, remix lineage, and licensing terms that persist even as content is copied, shared, or modified across learning platforms. This improves academic integrity, supports reproducible research, and fosters ethical reuse.
Creation Rights is more than a metadata system, it is the rights infrastructure for the next generation of digital creativity, enabling compliance-aware ecosystems where creators are compensated, AI is aligned, and innovation flows freely.
Creation Rights offers the equivalent of TCP/IP for licensing: a foundational protocol layer that travels with media, updates as it's used, and speaks the same language across platforms. Metadata is not merely supplemental to content, but constitutive of its legitimacy, provenance, and value. Much like TCP/IP serves as the foundational protocol for the transmission of data across the web, Creation Rights aims to be the foundational protocol for transmitting and enforcing intellectual property rights. In this future, every digital asset (whether a video, image, AI model, dataset, or creative work) carries with it an immutable and machine-readable declaration of its terms, identity, and usage conditions.
Metadata fabric is more than tagging or indexing, it represents an interoperable rights infrastructure that seamlessly binds identity, attribution, licensing, lineage, and value to every piece of digital content. It’s a new internet substrate that is inherently aware of ownership, compliance, and compensation, not bolted on as an afterthought allowing for creative consumption. The dynamic fabric travels with content across platforms, devices, protocols, and even blockchains, enabling verifiable use in AI training, content distribution, and derivative creation without ever losing the connection to the creator.
Creators gain power and visibility, and their rights are encoded into their files, their contributions are tracked through remix chains, and their earnings are routed automatically through smart royalty systems. Audiences gain trust in authenticity, knowing that they are seeing, hearing, and engaging with media that has a verifiable origin. Platforms gain a compliance and licensing layer that improves integrity and reduces legal friction.
AI becomes not a threat to rights, but a co-creator that respects them. Generative systems can query Creation Rights metadata to determine what content is legal to use, what rights must be attributed, and how to route licensing or royalties back to creators. This prevents unauthorized scraping, supports opt-in training pools, and creates a sustainable incentive loop between AI models and content creators.
Creation Rights is a programmable rights layer designed for co-creation, both human and AI-generated content. It embeds fair attribution directly into each file, making sure creators are always credited. It provides verifiable licensing that can be read and enforced by any machine. It supports remix-aware lineage, enabling transparent tracking and credit for creative evolution. It is built with a monetization-first design, tailored to support creative consumption and the creator economy. We extend an invitation to the entire ecosystem: we invite developers to build on our open APIs. We invite creators and AI companies to join our protocol. We encourage platforms to integrate metadata recognition for a more equitable digital landscape. And we call on policymakers to champion interoperable metadata standards. As content becomes increasingly dynamic, its metadata must evolve too, making ©® native to the internet.
Contact
team@creationrights.ai
creationrights.ai
Copyright ©® 2025 creationrights.ai
Version: 1.0
Date: June 2025
Prepared by: Mr. F. Elusio Zafar and Jeff Gordon (Founders of Creation Rights)
References
[1] John Locke, Two Treatises of Government, 1690.
[2] Adam Smith, The Theory of Moral Sentiments, 1759.
[3] Adam Smith, An Inquiry into the Nature and Causes of the Wealth of Nations, 1776.
[4] Jeremy Bentham, An Introduction to the Principles of Morals and Legislation, 1789.
[5] David Ricardo, On the Principles of Political Economy and Taxation, 1817.
[6] John Stuart Mill, On Liberty, 1859.
[7] Émile Durkheim, The Division of Labour in Society, 1893.
[8] Thorstein Veblen, The Theory of the Leisure Class, 1899.
[9] Ronald Coase, The Nature of the Firm, 1937.
[10] Joseph Schumpeter, Capitalism, Socialism and Democracy, 1942.
[11] Marshall McLuhan, The Medium is the Message, 1964.
[12] Roland Barthes, Death of the Author, 1967.
[13] Pierre Bourdieu, Distinction (A Social Critique of the Judgement of Taste), 1986.
[14] Elinor Ostrom, Governing the Commons (The Evolution of Institutions for Collective Action), 1990.
[15] Yochai Benkler, The Wealth of Networks (How Social Production Transforms Markets and Freedom), 2006.
[16] Lawrence Lessig, Remix (Making Art and Commerce Thrive in the Hybrid Economy), 2008.
[17] Simon Sinek, The Infinite Game, 2019.
[18] Jessica Silbey, Against Progress (Intellectual Property and Fundamental Values in the Internet Age), 2022.
[19] Rick Rubin, The Creative Act (A Way of Being), 2023.
[20] Fawad Zafar, Systems and methods for generating dynamically updated metadata using real-time artificial intelligence models. https://patentimages.storage.googleapis.com/16/14/3b/9fb63e3c90fb70/US11816474.pdf, 2023.
[21] Carolyn Dailey, The Creative Entrepreneur (A Guide to Building a Successful Creative Business from Industry Titans), 2025.
PDF download: https://drive.google.com/file/d/1lTrOXJLvtMTRHw3avZ-BIsm-56CMEt7X/view?usp=sharing
