Most AI-first labels hide. When a brand legal team asks the obvious questions, they don't have the answers, or won't share them. San Carlos Records picks the opposite path. The research is part of the product. Every doctrinal position we hold is published — in plain English, with the legal-eye analysis a click away, with a fresh "verified" date stamped on every page.
The tools are new. The law is moving. The platforms are inconsistent. And the honest answers — about who owns what, about which disclosures are required where, about what survives a brand's legal review — are uncomfortable. Most labels reckon that the path of least resistance is to avoid the conversation. So they publish a logo, a few generated tracks, a checkout, and a vague claim that "you own the output."
That's not malicious. It's just the cheap path.
San Carlos Records publishes the doctrine instead. Every claim about who owns what is grounded in a research brief you can read. Every track ships with a provenance bundle that documents how it was made. Every quarter, a routine watches our provider terms of service, stamps a fresh "verified" date on the artifacts you're looking at, and writes a statement of implication if anything material has changed.
If you're going to buy music from an AI-first label, you should be able to read the label's working assumptions before you commit. That's what this page is for.
Almost every important question about AI music collapses into two separate doctrines. Most label communication blurs them; we separate them deliberately so you can see exactly what you're getting.
Did a human meaningfully author this track?
Set by the US Copyright Office's January 2025 framework. AI is treated as an instrument; human direction is the authorship; mere prompting doesn't qualify. Our five-pass production methodology — lyric drafts before prompts, prompt construction as a documented decision, generation as one step among many, mixing and arrangement as human-direction — is purpose-built to clear this doctrine the way the Copyright Office rewards.
Should this track be flagged as AI-generated where it's uploaded?
Asked by SoundCloud, Spotify, YouTube, Apple Music, and TikTok independently — each with its own rule, threshold language, and enforcement posture. SoundCloud reads broadly; Spotify reads narrowly; YouTube focuses on deepfake concerns; Apple uses editorial discretion; TikTok asks broadly. The doctrine is engineering, not philosophy: we know which interventions move a track across each threshold, and we publish the matrix.
The two doctrines clear or fail independently. A track can be copyrightable (yours to license) while still requiring a platform disclosure flag (because AI audio is in the master). It can also satisfy disclosure-relief without enough human direction to be copyrightable. We design our tiers around the difference.
Each tier is a clear statement: what we deliver, what you get, what it means for your use, and what action you need to take. Read the one that matches what you're commissioning. Each links to the full per-tier customer contract page with the legal detail.
This is the engineering doctrine that drives our per-track disclosure decisions. The full per-platform analysis lives in research/11 §4. The agents enforce it at upload-gate.
| Pipeline pattern | SoundCloud | Spotify | YouTube | Apple Music | TikTok |
|---|---|---|---|---|---|
| Full AI (Suno master, AI mastered) | Disclose | Disclose | Disclose | Editorial risk | Disclose |
| AI bed + live vocals | Disclose | No disclose if AUTHORLOG complete |
No disclose | Editorial OK | Disclose |
| AI bed + live vocals + live overdub | Disclose | No disclose | No disclose | Editorial positive | No disclose if AUTHORLOG complete |
| AI in pre-production only (no AI audio in master) | No disclose | No disclose | No disclose | Editorial positive | No disclose |
T1 typically sits in row 1 or 2. T2 in row 2 or 3, with per-platform decisions per AUTHORLOG state. T3 is engineered to sit in row 4 — the arbitrage zone where the dual clearance is genuinely available.
Every claim above is anchored in an artifact you can read. Most are linked here; the per-customer documents (AUTHORLOG, Provenance Statement) ship with each commission.
Per-track record of every human intervention — studio sessions, vocalists named, instruments overdubbed, stems replaced, mixing decisions. Ships with every T2 and T3 commission. The audit defense.
Every released track carries an Adobe Content Credentials cryptographic provenance manifest. Adobe is the signer; we never self-sign. Verify on your side that what you received is what we sent.
Every SKU we sell has a row documenting upstream rights, named characters, source materials, decisions. Nothing ships without a clean clearance row.
A monthly routine reads each provider's terms of service and blog (Suno, Udio, ElevenLabs) plus the Sony-v-Suno docket. Changes trigger an implication statement, update affected readouts, and stamp a new verified date on this page.
The underlying legal-eye briefs. Written for a working legal reader, footnoted, grounding every claim on this page.
Per-commission signed document restating the tier, chain-of-title, disclosure flags, and license scope. Pinned to the doctrine version live at your commission date — your snapshot is yours forever, even as doctrine evolves.
Edge cases haven't been thought through yet, and we'd rather work them out with you before contract than after invoice.
A 60-second brief. Tier, intent, budget envelope. We reply within two business days with a quote, contract draft, and a doctrine snapshot pinned to today's date.
Not yet sure which tier? Re-read the T1, T2, or T3 customer statements above. Or describe the use case in the brief below and we'll route to the right tier.
It does not replace a contract. The per-tier pages tell you what you're buying; the actual commission contract is the binding document and goes through SCR's legal review at order time.
It does not replace your own legal review. The doctrine here is what SCR holds and publishes. Your jurisdiction may add overlays we don't track (we cover US platform doctrine and acknowledge EU AI Act phasing in the underlying research). If your use is high-stakes, run it past your counsel.
It does not freeze. The published doctrine moves with the upstream landscape. Provider ToS shift. Platform policies tighten. Court rulings reshape what's defensible. The version history on this page is the audit trail for every change we've made. If you commissioned a track at v1.0 and we're now at v1.4, your bundle is still v1.0 — your contract is pinned to the doctrine version live at your commission date.