Complete decision surface — RightsDocket taxonomy engine
2
Provenance tiers
(Gold / Silver)
6
Creation scenarios
+ 1 blank path
11
Asset types
at intake
56+
Discrete binary
determinations
Layer 1 — Provenance tier (first decision)
Gold — Secure Staging
Real-time capture during creation
RFC 3161 trusted timestamp anchored at creation moment. C2PA manifest signed and embedded. Proves both what was made and when with cryptographic chain.
Coming soon — real-time capture with cryptographic chain of custody
Silver — Retroactive
After-the-fact documentation
Same wizard, same claim outputs. No real-time timestamp anchor. Proves what was made and by whom — temporal chain requires external evidence.
Available now — document existing work with structured claim preparation
Layer 2 — Creation scenario (7 paths, pre-mapped claim language)
| # | Scenario | Claim basis | Exclusion basis | Risk |
|---|---|---|---|---|
| 1 | AI instrumental + human vocals, lyrics, melody | Lyrics, vocal melody, vocal performance, arrangement of stems | AI-generated beats, backing tracks, stems | Medium |
| 2 | AI melody/chords + human arrangement, lyrics, performance | Revised melody, lyrics, arrangement, performance | Unmodified AI melody + chord progressions | High |
| 3 | AI-assisted production only (de minimis) | Full musical work + sound recording (all human) | Nothing — no Limitation of Claim needed | Low |
| 4 | Substantially reworked AI demo | Rewritten lyrics, revised melody, restructured arrangement, new performances | Underlying AI demo + any unmodified AI expression | High |
| 5 | Human composition + AI-generated cover art only | Full musical work + sound recording (all human) | AI-generated cover artwork / promotional visuals | Low |
| 6 | Multiple AI tools across different elements | Selected/edited lyrics, arrangement choices, vocal performance, production direction | AI stems, unmodified AI lyrics, AI mastering, each tool itemised | High |
| 7 | Start from scratch (blank wizard) | User-defined | User-defined | — |
Layer 3 — Musical element taxonomy (what can be claimed, flagged, or excluded)
Musical work (Form PA) — claimable
Music (checkbox)
Top-level USCO selection
Lyrics (checkbox)
Cleanest human claim
Vocal melody
Subcategory of Music
Chord progressions
Only if human-authored; standard progressions not protectable
Harmony
Standard sequences (C,F,G,C) unprotectable per Compendium §800
Melody
Highest risk if AI-generated; core compositional element
Arrangement / song structure
Verse/chorus/bridge, sectional design, instrumentation decisions
Selection & coordination of AI stems
Must be "sufficiently creative" — Zarya of the Dawn standard
Rhythmic structure
Human decision, not AI default grid
Lyric revisions / edits to AI text
Must meet Feist originality bar — not just light editing
Sound recording (Form SR) — claimable
Sound Recording (checkbox)
Top-level USCO selection
Vocal performance
Clearest human claim
Instrumental performance
If live / human-played
Production direction
Creative decisions made in session
Stem selection & sequencing
Must document specific choices made vs. AI defaults
Editing & mixing decisions
Must be creative, not purely technical
AI mastering
De minimis / technical processing — not authorship
AI pitch correction / auto-tune
Technical production, not expressive authorship
AI stem separation
Technical, not creative
Material excluded — must be disclaimed
AI-generated instrumental stems
Core exclusion for Suno / Udio workflows
AI-generated beats / backing tracks
Prompt-to-audio generation
AI-generated vocal synthesis
Even if lyrics were human-written
AI-generated melody (unmodified)
Modified versions may be partly claimable
AI-generated chord progressions (unmodified)
Same as melody
AI-generated lyrics (unmodified)
Edited versions may qualify; raw output does not
AI mastering / stem processing
Exclude if > de minimis; omit if purely technical
AI cover art / album artwork
Excluded from musical work claim
Previously published material
Standard Limitation of Claim requirement
Third-party owned material
Platform ToS stems — platform may retain license
Layer 4 — All evaluated decision categories (full count)
11
Asset types at intake
Song/Single · EP Track · Album Track · Instrumental · Beat/Loop · Jingle · Ad Cue · Brand Theme · Soundtrack Cue · Remix · Other
7
Creation scenario paths
6 pre-mapped + 1 blank. Each triggers different claim template, field pre-fill, and risk flag logic
10
Musical work elements evaluated
Music · Lyrics · Vocal melody · Chords · Harmony · Melody · Arrangement · Selection · Rhythm · Lyric edits
9
Sound recording elements evaluated
SR checkbox · Vocal perf · Instrumental perf · Production · Stem selection · Editing · Mastering · Pitch correction · Stem separation
10
Exclusion categories evaluated
AI stems · Beats · Vocal synth · AI melody · AI chords · AI lyrics · Mastering · Cover art · Prior publication · Third-party owned
5
Readiness dimensions scored
Contributor docs · AI disclosure · Human specificity · Separability · Evidence attachments
3
eCO form fields generated
Author Created · Material Excluded (Other) · Note to Copyright Office
4
Filing pathway options
Standard App ($65) · Single App ($45) · GRUW (up to 10) · GRAM (up to 20, $65)
3
Copyright object types
Musical work (Form PA) · Sound recording (Form SR) · Both on one Standard Application
3
Legal precedents coded
Zarya of the Dawn · Rose Enigma (separability) · Feist Publications (originality)
3
Evidence attachment categories
Audio (master/stems) · Session/source files (MIDI, DAW) · Documents (lyrics, contracts, screenshots)
2
Provenance tiers
Gold (Secure Staging — real-time RFC 3161 + C2PA) · Silver (Retroactive — documented after creation)
56+
discrete determinations per claim
Every RightsDocket claim traverses 2 provenance tiers, 7 scenario paths, 10 musical work elements, 9 sound recording elements, 10 exclusion categories, 5 readiness dimensions, 4 filing pathways, 3 copyright object types, and 3 legal precedents — generating 3 eCO-ready form fields via deterministic rule mapping, not AI generation.The eCO form has 3 fields and 2 checkboxes on the Limitation of Claim screen. RightsDocket has 56+ decision nodes behind them.
Not LLM-generated claim language — deterministic rule mapping from user inputs