Casset Canonical Thesis
Status: Canonical strategy source · Owner: Founder · Last updated: 2026-06-24
One-Line Thesis
Casset is the artist home for audiovisual identity in the generative media era: humans enter cinematic Profile Worlds, while agents read canonical release context, provenance, permissions, and lineage.
Why Now
Music is entering an abundance phase. Audio, visuals, captions, remixes, and release assets can be generated faster than culture can interpret them. Distribution platforms still flatten releases into inventory, and generative systems can multiply content without carrying authorship, consent, atmosphere, or context forward.
The scarce layer is no longer only the file. It is:
- artist identity
- authored audiovisual context
- provenance
- contributor intent
- permission semantics
- release lineage
- social memory
- taste
Casset exists to be that layer. It is the place an artist's music world becomes emotionally legible to people and structurally legible to machines.
Category
Casset is an audiovisual music identity platform with an emerging canonical release layer underneath it.
As the product matures, that same infrastructure layer generalizes into an operating system for creative assets. The release layer already treats a published work as a canonical, machine-readable object; the natural extension is that every creative object on the way to a release — a track, a sample, an instrument, a visual — carries the same identity, ownership, provenance, and permission posture across its whole lifecycle. The audiovisual identity is still what people feel, and the Release is still the canonical published artifact; the asset layer simply keeps the work legible before, during, and after that moment. This is an expansion of the existing thesis, not a new direction — Profile Worlds, the four primitives, and the canonical release layer are unchanged.
The public product remains:
- Profile Worlds
- Hook Objects
- Release Rituals
- Listening Rooms
- atmosphere/world authoring
- cinematic playback
- social release memory
The infrastructure layer quietly adds:
- release manifests
- provenance events
- contributor graphs
- permission policies
- agent access policies
- derivative lineage
- release hashes
- optional Base anchoring for proof
Casset is not:
- a streaming app
- a generic music social network
- an AI music generator
- a generated-content firehose
- a creator monetization SaaS
- a rights-management dashboard
- a crypto product
- a creator coin or market platform
- a smart-link or pre-save utility
Core Loop
text
human loop:
enter Profile World
-> play Hook Object
-> feel Atmosphere
-> join the Listening Room
-> participate in the Release Ritual
-> return to the artist home
machine loop:
resolve release
-> read canonical manifest
-> evaluate permissions
-> inspect provenance
-> traverse lineage
-> request access or register derivative
The two loops should not collapse into one UI. Humans should experience music and identity. Machines should consume structured release context.
Artist Home
The Profile World is the artist's home on Casset. It is not a static page, a link hub, or an admin wrapper around tracks.
It should answer:
- What does this artist sound like?
- What world does the song live inside?
- What should listeners remember?
- What is authored, assisted, remixed, or generated?
- What release moment is happening now?
- What can agents or systems do with this release?
The product surface should make that feel emotional. The infrastructure should make it durable.
Four Product Primitives
Profile World
The living audiovisual identity for a creator, fan, duo, group, label, estate, or scene. It contains sound fragments, atmosphere, visual language, social presence, provenance signals, collaborators, follow state, current rituals, and release context.
Hook Object
The smallest emotional audiovisual unit: an audio window plus timed lyrics or caption text, waveform state, visual world, theme tokens, social proof, provenance, share state, reply/remix affordances, and links back to the canonical artist home.
Release Ritual
A time-bound social moment around a song or hook: pre-release rooms, early supporters, presaves, fan clips, co-cassets, unlocks, comments, participation proof, and post-release memory.
Listening Room
The presence layer around a profile, hook, casset, or release. It is not generic chat. It should feel like people gathering inside the emotional weather of a song.
Emerging Infrastructure Primitives
These do not replace the four product primitives. They make them durable in a generative, agent-mediated media ecosystem.
Release Object
The canonical release definition behind a public music experience. It connects an artist, tracks, versions, contributors, permissions, provenance, lineage, and published state.
Release Manifest
A deterministic machine-readable snapshot of a release version: identity, tracks, audiovisual refs, contributors, splits, permissions, and schema version. Agents consume manifests; listeners consume the cinematic surface.
Permission Policy
Artist-authored or default release rules for remixing, stems, sync, visual reuse, AI training, AI generation, derivative registration, commercial usage, and licensing routes.
Provenance Event
An append-only record of release creation, versioning, manifest generation, signing, anchoring, contributor updates, permission changes, and derivative registration.
Agent Access Policy
A structured answer to the question: what can this agent do with this release, under which terms, and what path exists if a license is required?
Creative Asset And Asset Reference
A creative asset is any addressable object an artist makes — track, sample, instrument, visual, or release. An Asset Reference (AssetRef) is the deterministic, typed handle to one: its asset type, id, owner, content hash, a short-lived resource link, a resolved permission decision, and a back-pointer into provenance and lineage. AssetRef is the contract that tools, creative runtimes, and agents consume across the asset lifecycle — before, during, and after release. It does not hold rights itself; it carries a permission decision resolved from the release layer.
Creative Asset Layer
This section generalizes the canonical release layer; it does not replace it. The release layer already makes a published work canonical and machine-readable. The creative asset layer extends that same discipline backward and forward across the lifecycle, so the work stays legible before and after it becomes a Release. It is stated conservatively, so it never destabilizes the identity product above it or the release layer it grows from.
Lifecycle. Assets exist before, during, and after release. A sample uploaded as raw material, a generation made from a track, an instrument opened in a creative runtime, a remix produced by one — these are all assets long before any of them becomes a published Release.
What stays canonical. The Release remains the canonical public/published object. The ReleaseManifest remains the canonical machine-readable contract for published releases. AssetRef is the broader contract for creative objects across the whole lifecycle — it complements the manifest, it does not replace it.
Asset types, not one model. Track, Sample, Instrument, Visual, and Release are asset types. Samples are one asset type among several, never the canonical model. A thin AssetRef projects over all of them; their underlying models stay distinct.
Rights stay in the release layer. Rights, permissions, and commercial provenance remain anchored in the Release / PermissionPolicy / DerivativeLink / ProvenanceEvent layer. AssetRef carries a resolved permission decision (allowed, denied, contact owner, license required); it is not itself a rights record.
Creative runtimes are agents. External creative tools — the kind that open an asset, transform it, and produce a new one — are a kind of agent. They consume AssetRefs and produce new assets, and they resolve permission through the same engine that answers agent permission checks. Casset owns the identity and the lineage; the runtime owns the execution. The architecture should make room for many creative runtimes and privilege no single implementation. Any specific tool is one consumer of the asset layer, never part of the thesis.
Two lineage histories, one bridge. The creative lineage graph records what derived from what across all asset types (discovery and attribution). The rights graph records permissioned derivatives between releases. The creative graph projects into the rights graph when a derivative becomes a published Release, so the two histories never drift apart.
Discipline. The asset layer earns its place only by deepening the four primitives and the canonical release layer. It is not a generic file manager, a digital-asset-management product, or a marketplace. If an asset feature does not strengthen a Profile World, Hook Object, Release Ritual, Listening Room, or the release layer, it parks.
Generative Content Stance
Casset is not anti-AI and not AI-utopian.
The risk is not that generated media exists. The risk is generated media with no home, no taste, no consent, no provenance, no artist context, and no permission surface for systems that need to act responsibly.
AI can:
- lower friction for visual worlds, lyric timing, hook selection, captions, and
atmosphere variations
- help non-technical creators shape audiovisual identity
- support hybrid creators with transparent provenance
- summarize fan traces or release momentum for artists
- help agents discover allowed uses without scraping intent from prose
AI should not:
- publish without user taste and intent
- reward high-volume low-context output
- erase provenance
- blur permission boundaries
- make the public product feel like content automation
Public language should be nuanced: human-made, no AI used, AI-assisted, generated visual world, remix, co-created, and derivative are provenance signals, not moral rankings.
Agentic Permissions
As agents begin to search, summarize, remix, license, train on, and route media, they need structured permission answers. Casset should become the place where an artist can express those answers once and have them travel with the release.
The near-term permission surface is not a legal operating system. It is a clear, machine-readable policy layer:
- allowed
- denied
- contact owner
- license required
The strategic direction is:
text
agent asks: may I use this release for X?
-> Casset resolves the canonical release
-> Casset evaluates the permission policy
-> Casset returns decision, reasons, and route
-> license, access, or derivative registration can happen without losing provenance
This is artist infrastructure, not a dashboard aesthetic.
The same permission surface serves creative runtimes that open assets. A runtime that wants to load an asset gets the same allowed / denied / contact owner / license required answer an agent would, resolved by the same engine. Permission is expressed once and travels with the asset, whether the consumer is an autonomous agent or a tool a human is driving.
Base Stance
Base is not the product.
Base can support Casset by anchoring release manifest hashes, provenance proofs, settlement references, and future licensing receipts. The listener should not need to understand Base, wallets, markets, or token mechanics to experience a release.
Use Base language only as infrastructure language:
- anchored proof
- manifest hash
- provenance anchor
- receipt
- settlement reference
Avoid:
- creator coin
- token tab
- market mechanics
- trading surface
- speculative ownership language
Visual Studio Stance
Visual Studio is an atmosphere/world-authoring system for songs.
It should feel like choosing the emotional weather of a hook: source media, profile texture, color, lyric behavior, motion, and rendered artifacts. Shader or VFX language can describe internal implementation, but it should not be the primary product framing.
Optional And Peripheral Systems
Commerce, campaigns, drops, rewards, analytics, protocol experiments, and settlement primitives can remain in the platform, but they are not the category. They matter only when they deepen a Profile World, Hook Object, Release Ritual, Listening Room, or canonical release layer.
Implementation posture for support, experimental, and parked systems lives in docs/architecture/peripheral-systems.md.
Language To Prefer
- audiovisual music identity platform
- artist home
- Profile World
- Hook Object
- Release Ritual
- Listening Room
- atmosphere
- profile texture
- audiovisual object
- visual world
- canonical release context
- release manifest
- provenance-aware identity
- agent-readable permissions
- release lineage
- human, no-AI, AI-assisted, generated, remix, or co-created provenance
- operating system for creative assets
- creative asset
- asset reference (AssetRef)
- asset lifecycle
- creative runtime
- creative lineage graph
Language To Avoid As The Lead
- social network for sound identities
- AI-powered music platform
- AI music generator
- shader editor
- VFX tool
- growth platform
- campaign dashboard
- tokenized music platform
- creator coin platform
- blockchain product
- pre-save optimization tool
- fan monetization SaaS
Emotional Center
For creators: "This is the home for my music identity, release context, and permissioned future."
For listeners: "I understood the person and the song more deeply because I was inside the world around it."
For agents and systems: "This release has a canonical source, readable permissions, provenance, and lineage."
For culture: "Intentional music kept its identity as generation became abundant."