An invitation

Become a Verified Voice.

We're inviting a small bench of practitioners to co-build verified versions of themselves — to put the kind of sparring that today happens in boardrooms in front of a kid in Rio, a student in a 6m² room, a founder up at 2am. Free for them. Profit-shared with you. Contract-backed. Reversible at any time.

01

Why you

We're not looking for influencers. We're looking for practitioners with a defensible point of view — people who've done the work, taken the scars, and articulated something the field actually argues with. Founders, operators, scientists, judges, doctors, comedians, athletes, craftspeople, dissenters, teachers.

  • — A body of public work (essays, talks, books, posts, decisions, products).
  • — A stance you'll defend out loud, not a personal brand.
  • — Comfortable being sparring-partnered at scale, with editorial sign-off.
  • — Curious about what your thinking does in 100,000 unfamiliar hands.
02

Why us

THE ROAST is a critical-thinking sparring engine. Our Charter is one rule: critique ideas, never identities. Verified Voices sit alongside historical figures (Sun Tzu, more coming) and a small council of archetypes — all engineered, all ledgered, all reversible.

Why we exist is on /why: quality thinking and quality sparring partners shouldn't be a postcode lottery. We're paying it forward — and we want a few sharp minds to come with us.

03

Why together

One verified voice, well-engineered, can sit in roughly ~10,000 free critical-thinking sessions a year at our current cost basis. That's not a marketing line — that's the unit economics of putting your best 30-minute sparring at the disposal of people who'll never get on your calendar.

For you: distribution, a compounding canon (every signed-off answer becomes part of your voice pack, not ours), and a paid tier you share in. For us: a defensible bench. For the people we're paying it forward to: a brain in the room they wouldn't otherwise have.

04

What you contribute

🎙️
One discovery interview
~2 hours, recorded. We use it to fill the Bones sheet (Want / Need / Wound / Lie / Ghost) and capture cadence.
📚
Media DNA
At least 30 influence fragments — essays, talks, books, decisions — each tagged with the trait it carries.
Sign-off rounds
You review the system prompt, sample transcripts and the first 20 canonicals before launch. Nothing ships without you.
📞
Quarterly drift check
30 minutes a quarter. Confirm we're still in your voice, or steer us back. Re-affirm or revoke consent — your choice.
05

What you give the community

The free tier is the mission. It's the kid in the slum in Rio with a fun idea. The student eating noodles in a 6m² room. The starter at the golf club who needs a different POV. The second-time founder at 2am who can't afford a board.

Your free-tier conversations are not revenue-shared — that's how we keep the door open for everyone. Paid tiers and enterprise are where the share kicks in.

06

How you earn money

✍️
Signing fee
Paid on counter-signature of the LL-1 contract. Amount set per voice based on canon depth and exclusivity tier you choose.
📈
Revenue share
A percentage of net Voice-Pack-attributable revenue from paid tiers and enterprise seats. Per-session attribution, audited yearly, transparent dashboard.
🗓️
Quarterly payouts
Cycles close Mar 31 / Jun 30 / Sep 30 / Dec 31. Payouts within 30 days of close, with an itemised report.
🔍
Audit right
Once a year, your nominated auditor can inspect usage logs under NDA. We open the books.
07

Worked example — what does this actually pay?

Numbers below are illustrative only. Final session price, share % and signing fee are agreed per voice at signature. Historical voices (e.g. Sun Tzu) don't trigger revenue share — only Verified (real) Voices do.

ScenarioPaid sessions / moAvg session valueYour share %Quarterly payout
Conservative~500$TBDTBD%$TBD
Base~2,500$TBDTBD%$TBD
Strong~10,000$TBDTBD%$TBD

We'll fill these in with you on the vetting call so the deal is never abstract — you'll see the math before you sign.

08

Your safety, due diligence, your rights

The full agreement is the LL-1 Likeness, Name & Voice-Pack License (LL-1-2026-05). Plain-English summary of what you get out of it:

🛑
7-day kill switch
Any reason, no justification. We pull the public Voice Pack within 7 days of your written request. Already-shared outputs aren't retroactively nuked from people's screens — but no new ones are served.
🚫
Deny-list
Categories your voice will never enter: politics, medical, defamation, sexual contexts. Published. You can extend it.
Moral correction right
Demand take-down of any specific output. We comply or explain in writing within 7 days.
🔁
Quarterly re-affirmation
Every 90 days you confirm or revoke from your verified email. Consent is never assumed; the snapshot hash is recorded.
🏷️
Exclusivity is your call
Pick exclusive, first-look, or non-exclusive at signature. We don't pressure-pick for you.
🤝
Mutual scandal clause
Either side can pause if a public scandal materially harms the other, pending good-faith review.
🔐
Data & privacy
Source materials encrypted at rest, admin-gated, GDPR/CCPA handled. User conversations are never used to train models.
🧹
Clean exit
On termination: public Voice Pack down in 7 days, interview recordings deleted in 30, final usage report shared with you.

Plain-English summary. The signed LL-1 governs. Take it to your own counsel — we'll wait.

09

The journey

  1. 01
    Apply
    5–10 minute form. We read every one personally.
  2. 02
    Vetting call
    30 minutes with us + an existing voice. Mutual fit, real pricing math.
  3. 03
    Sign LL-1
    Counter-signed contract. Signing fee paid. Schedule A inventory agreed.
  4. 04
    Discovery interview
    ~2 hours, recorded. Bones, cadence, do/don't list filled.
  5. 05
    Build & evaluate
    Corpus, canonicals, evals, ghost-delta < 0.25 vs baseline.
  6. 06
    Soft launch
    50-user cohort, daily drift monitor for 7 days.
  7. 07
    Public launch
    Live on the gallery. First quarterly payout cycle starts.

Typical timeline from signed LL-1 to soft launch: 4–6 weeks.

10

FAQ

Do I keep ownership of my name and likeness?
Yes. You license them narrowly to us for running the Voice Pack. The licence ends the moment you ask it to.
What if I become controversial mid-term?
Mutual scandal clause: either side can pause pending good-faith review. You're never trapped.
Can my estate take over after I'm gone?
Yes — the LL-1 includes a survivorship section. We can also transition you to a 'historical voice' framework with citation ledgering.
Who can see what people ask my voice?
You can, on request, under NDA. Aggregate analytics are in your dashboard. We never sell raw conversations.
Do you train models on user inputs?
No. User inputs to your persona are not used to train any model — yours, ours, or anyone else's.
How are payouts taxed?
Paid as royalties to your nominated entity. We provide annual statements; tax treatment is yours to determine with counsel.
Can I quit and take my canon with me?
You retain rights to your underlying works. The compiled Voice Pack engineering (prompts, evals, code) stays with us. The canon of signed-off responses is jointly attributable; you get a full export on exit.

If this sounds like the right room — apply.

8-minute form. We reply within 7 days, every time.

Apply to be a Voice →