Ritual Creative Passport

Can we publish this?

Every asset Ritual delivers carries the answer. One signed, machine-readable record of rights, models, claims, disclosure and approval, built inside the production pipeline so compliance happens as the work is made.

AI-assisted. Human-directed. Brand-controlled.

Creative PassportSample record
Asset IDRTL-2026-0912-A7F3
Content hashsha256 9f2c41…e41b
ProjectAurora Sparkling, Holiday :30
AI involvementGenerated, synthetic performer
TerritoriesUS-CA, US, EU
Signed2026-09-12T18:04:22Z
  • Brand approved
  • Human reviewed
  • Source assets licensed
  • Likeness and voice rights verified
  • Claims substantiated
  • AI models documented
  • Required disclosure applied
  • Content Credentials embedded
  • Territory approved
  • Final approver recorded
Cleared
for export

Illustrative record for a fictional brand.

Any model can make the story. Only the record makes it trustworthy.

April 2026C2PA 2.4 adds a machine-readable AI Disclosure assertion to Content Credentials.
August 2, 2026The EU AI Act's Article 50 transparency duties and California's AI Transparency Act become operative on the same day.
September 16, 2026California signs SB 1050, requiring explicit disclosure when an ad uses an AI-generated performer to sell.
The record

A permanent passport for every asset.

The Passport is the RITUAL record made visible. Six fields, in the order they occur in the life of a story, carrying everything a brand, a platform or a regulator will ask about a finished piece of creative.

R
RightsWho agreed?
source assetsownership and license statuslikeness rightsvoice rightstraining permission, held separately
I
IntentWhat was it for?
clientprojectbriefclaims permittedclaim substantiation
T
TraceWhere did it come from?
asset ID and hashevery model, exact versiongeneration and edit timestampsgenerated, manipulated or assistedhuman creators and editorsupstream credentials preserved
U
UseWhat is it allowed to be?
territory restrictionsdisclosure decisionvisible label appliedfinal human approverexport history
A
AttributionWhat did it do?
channels deliveredperformance bound to the creative decision
L
LearningWhat does the next one begin knowing?
write-back to the brand's own intelligencenever to a general model without contractual opt-in

Sealed with C2PA Content Credentials as the default provenance standard. We treat C2PA as an implementation standard, never as a legal safe harbor.

The specification behind the record
The gate

Unresolved means unexportable.

Each project carries its destinations and its flags. The disclosure engine decides which label, metadata and approvals the work requires, and export stays locked while a mandatory flag is open. Try it.

AI involvement
Destination
Project flags
On file
Export blocked
Disclosure level
Visible label
Machine-readable
Open flags

    Illustrative logic. Label wording follows the governing law wherever a statute prescribes it, and the production rules table is validated by counsel for each jurisdiction before launch. Customers cannot switch off a legally required disclosure.

    Disclosure policy

    Three levels. Never a blanket label.

    A binary label punishes honest craft and hides real risk. Ritual records all AI use, and escalates what the audience sees in proportion to what could deceive them.

    Level one

    AI-assisted craft

    Cleanup, rotoscoping, color, minor editing, transcription, ideation.

    Ritual defaultNo blanket visible label. AI use recorded in the Passport and the Content Credentials.
    Level two

    Materially AI-generated

    Generated picture or sound with no realistic human deception.

    Ritual defaultProvenance embedded. Visible disclosure set by jurisdiction, client and platform.
    Level three

    Synthetic people

    Realistic synthetic performer, digital replica, cloned voice, or an AI agent interacting as if human.

    Ritual defaultVisible disclosure, machine-readable provenance and documented rights. All three, every time.
    Synthetic fictional human"This performance features an AI-generated synthetic performer."
    Authorized real-person replica"This performance includes an AI-generated digital replica used with permission."
    Conversational agent"You're interacting with an AI assistant for [Brand]."
    Hard lines

    Rules the pipeline keeps, so nobody has to remember them.

    Provenance is a chain of custody

    When a model attaches Content Credentials or a watermark, Ritual preserves it and appends its own signed production step. Upstream provenance is never stripped.

    Likeness is a hard rights gate

    No export of a synthesized face, voice or performance of an identifiable person without documentation covering person or estate, use, duration, territory, media, modification and derivatives.

    Appearing in an ad is not consent to train

    Permission to use someone's image never implies permission to train or fine-tune a model on them. Training is a separate contractual grant.

    Every claim has a source

    "Clinically proven," "#1 brand," price, warranty, nutrition, sustainability. Each factual assertion links to an approved source in the Claim Ledger before export. A disclosure label does not cure a deceptive claim.

    Your data is not training data

    Client assets, prompts, campaign history and performance data do not train a general Ritual model or a third-party model unless you opt in by contract. Retention, deletion, encryption and access controls are standard.

    Only approved models are selectable

    Every generation vendor is reviewed for commercial rights, data retention, training-on-input, indemnity and provenance support. A model that cannot meet a job's disclosure duties cannot be chosen for that job. Model-agnostic by design.

    Sensitive work routes to counsel

    Political advertising, healthcare and financial claims, biometric or emotion analysis, minors, and replicas of recognizable people carry a Special Legal Review flag that sales cannot clear alone.

    A human signs last

    Every Passport names its final human approver. The record shows who directed the work, who edited it, and who released it.

    Machine-readable

    Built for legal teams. Readable by agents.

    The same record your counsel reviews is available to your systems. A media platform, a brand safety tool or an autonomous buying agent can ask one question before an asset runs, and get a signed answer.

    Enterprise-visible in the Ritual workspace, with API and MCP access to compliance status.

    GET /v1/assets/RTL-2026-0912-A7F3/passport200
    {
      "publishable": true,
      "territories": ["US-CA", "US", "EU"],
      "ai_involvement": "generated",
      "disclosure": {
        "level": 3,
        "visible_label": "applied",
        "content_credentials": "embedded",
        "assertion": "c2pa.ai-disclosure"
      },
      "rights": "verified",
      "claims": "substantiated",
      "final_approver": "recorded",
      "open_flags": []
    }
    Rolling out in this order, inside the production pipeline
    Asset Passport
    C2PA provenance
    Rights and consent ledger
    Jurisdiction disclosure engine
    Pre-export compliance gate
    Client-facing Creative Passport
    API and MCP access
    RRightswho agreed
    IIntentwhat for
    TTracewhere from
    UUseallowed to be
    AAttributionwhat it did
    LLearningwhat comes next
    The RITUAL Protocol

    The Passport is the product. The protocol is the standard.

    The RITUAL Protocol is Ritual's open, royalty-free specification for what must travel with a generated story. It wraps the C2PA manifest unchanged, so it adds to the provenance standard instead of competing with it. The Creative Passport is that specification running in production.

    Rights before generation. Use before delivery. Attribution before learning.

    Read the white paper

    Trustworthy enough to release.

    See a Passport travel with a finished spot, from first frame to final export.

    Request a walkthrough