Likeness¶
Pre-seed planning artifacts for a consent-first AI likeness platform for verified adult creators.
Verified creators license a private AI version of themselves to subscribers within rules they set, watermarked, revocable, with no model export. This site is the working corpus: the founder brief, the team plan, the budget, the design and validation work, the decks, and a clickable mockup.
Read in this order¶
The artifacts build on each other. New readers should follow this path:
- Founder Brief — what Likeness is, why it exists, what's defensible, what the hard parts are.
- Cofounder JDs — the six founding seats and what each one owns.
- Budget — how $1.5M becomes 18 months for a six-person team.
- ML Lead Technical Brief — the technical posture for the ML / Infrastructure cofounder seat.
- Design — Overview — the multi-phase startup-design pass: research, strategy, brand, product, financial, validation. Final scorecard and 30-day action plan.
- Decks — investor, cofounder, and creator decks rendered inline with PDF and PPTX downloads.
- Mockup — a clickable HTML mockup of the platform's creator and fan flows.
What's settled (don't relitigate without explicit reason)¶
These represent real choices made during planning. Treat them as defaults; if a request implies changing one, surface it rather than silently changing it.
Product¶
- Consent-first. Every output is watermarked, tied to a license, revocable. Models never leave the platform.
- No model export, ever. Architecture commitment, not just policy.
- No third-party reference uploads. Every face on the platform is identity-verified.
- License-gated inference. Every prompt parses against the creator's license object before any model is loaded.
- Per-output provenance. Watermark + perceptual hash + signed metadata on every generation.
- Per-creator adapter isolation. No bleed across creators by design.
MVP scope (concierge phase)¶
- 5–10 invite-only creators
- Manual model training (per-creator LoRA on a frozen open-weights base)
- Still images only at launch
- Subscriptions + compute credits via adult-friendly payment processor
- License engine + approval queue + watermarking + manual moderation
Explicitly not in MVP: video generation, voice cloning, public discovery / search, open API, fan self-insert, mobile app store distribution, unrestricted creator collaborations.
Founding team — six seats¶
- Founding CTO / Product (founder's seat) — owns stack, infrastructure, product direction, UX
- Founding CEO (recruit) — narrow scope: fundraising, investor relations, BD/partnerships, business strategy
- Compliance & Legal Lead — 2257 records, processor compliance posture, AI likeness law tracking, agreements
- ML Lead — model pipeline, license-gated inference rule engine, watermarking, face-matching for abuse
- Trust & Safety Lead — verification flows, classifiers, abuse vectors, enforcement, takedown pipeline
- Creator Relationships & Operations — adult industry insider; the bridge between platform and creators
Capital plan¶
- Pre-seed target: $1.5M for 18 months — calibrated for the six-seat founding team with ~11% reserve.
- Personnel-dominated budget (~62%). Compliance & Legal Lead is the most expensive single hire.
- Adult-friendly payment processors are the single most likely cause of needing a bridge round before seed.
What's not promised¶
- Perfect rights management (we say so, plainly)
- Recovery of files already downloaded
- Stopping screenshots
- Making gray-market sites disappear
About this site¶
Built with MkDocs using the Material theme. Source content is the markdown in this repo's docs/ and design/ directories, the decks rendered from dist/, and the mockup served from mockup/. The build pipeline converts the PowerPoint decks to PDF using LibreOffice in CI so they render inline on any device.
The repo is private; this published URL is unguessable rather than auth-gated. Don't share it more broadly than you intend it to circulate.