ManyKind.ai — The Consumer Flagship (Venture Deck)
SSCP · ManyKind.ai
Your AI. Your context.
Your wallet.
ManyKind.ai is the consumer flagship of the sovereign stack: a marketplace of
expert agents built on real coaches' methods, running against a data wallet the
user owns. It is the demand engine of the whole ecosystem: when ManyKind moves,
everything downstream moves.
Date Coach · live end to end
Hiring agent · first real users this week
Public launch · 2026-08-08
Creators keep their brand and clients
02 / 11Problem
The more useful your AI gets, the more of you it keeps
Consumer AI has one business model: absorb the user. And the expert
economy has one platform deal: give us your method, your clients, and most of the upside.
For people
- Your assistant's memory of you lives inside the vendor's perimeter. Switch and you start over.
- Every conversation trains someone else's product, in your own life's category.
- The most personal moments, the 11 PM spiral, the post-date debrief, the interview nerves, are exactly the data you least want absorbed.
For the experts people trust
- Coaches and specialists hold the scarce asset: judgment, packaged as a method. Platforms capture it.
- Platforms take a heavy cut and own the audience and the brand surface.
- Generic AI now answers their clients at 2 AM, with none of their method and none of their revenue.
Positioning source: memory bank 06-marketing/positioning/for-consumers.md.
03 / 11Why now
The owned consumer tier is open, and nobody has shipped it
Four forces converged: the trust break made people ask who keeps their context;
open weights got good enough for coach-shaped workloads; agents became capable of real work,
voice in, actions out; and the expert economy is actively looking for a way to package judgment
without surrendering it.
Trust
The frontier-API trust break reframed the consumer question from "how smart" to "who keeps what it knows about me."
Quality
Right-sized open models handle transcription, reasoning, and coaching reads. The workload fits the owned tier.
Capability
Voice-first intake, tool use, persistent state: agents can now do the work, not just chat about it.
Supply
Experts want distribution without capture. Their name on the door, their client book intact: a deal that recruits itself.
Your AI knows you. The question is: who owns that knowing?
memory bank · consumer positioning, core message
Sources: memory bank 00-foundations/vision-and-strategy.md (the four forces); 06-marketing/positioning/for-consumers.md (audience targets).
04 / 11Product
An entire consumer-AI company, in one place
ManyKind is an end-to-one solution: the agents, the models they run on, the data
they need, and the payments that meter them, orchestrated together under one standard of
security, autonomy, and accountability. More customized than generic AI, and cheaper,
because the whole loop runs on infrastructure the product does not have to rent blind.
Data wallet + vault
The user's context lives in a vault they own. Agents read and write only under signed, revocable grants. Every operation is signed and audited.
Expert-agent marketplace
Coaches publish agents built on their method, their brand front and center. Clients connect with a wallet; switching agents is a signature, not a migration.
Agent runtime + models
Voice intake, transcription, reasoning, and delivery run as orchestrated pipelines on open models, with state persisted in user-owned storage.
Accountability
Every operation on user data is signed and audited, and creator revenue is settled transparently.
05 / 11Why people believe it
Every fear maps to a shipped primitive
| The fear | What ManyKind ships against it |
|---|
| "It will keep my data" | Context lives in the user's vault: owned and revocable. Grants are checked before every data-touching operation and fail closed. |
| "I will be locked in" | Switching agents is one signature. Disconnect a coach, connect a better one, and it knows you, because your context never lived on the agent's side. |
| "AI will judge me" | Agents surface evidence, not verdicts: sourced observations with the exact moment attached. The human makes the call. |
| "The claims are inflated" | The website ships under an honest-claims rule: unshipped promises appear struck through, testimonials only when real and attributable. |
| "The expert gets exploited" | Creator-first economics: name on the door, client relationships stay with the coach. |
Sovereignty phrasing follows the honesty gate: owned and revocable today; deeper cryptographic guarantees ship on the roadmap and are not claimed early.
06 / 11Proof
Not a promise: a running loop
The full consumer loop, wallet, consent, streams, inference, state, payments,
is exercised by real agents today.
Date Coach · live end to end
Voice-first intake builds the profile. The user records a real date. The agent returns a structured coaching report: audio streamed into the vault, transcription and reasoning on cluster GPUs with open models, results in user-owned state, delivered through a vault-native widget.
flagship vertical · dating & relationshipsHiring agent · in real use
Given an interview recording, it surfaces the moments that matter with clickable audio evidence, playback, and downloads. First real recruiter ran live interviews through it this week. Same architecture, second vertical.
second vertical · hiring professionalsPlatform release
The unified platform release lets a non-engineer publish and run agents through the UI, sandbox to stage on the same APIs. If marketing can ship an agent through the flow, a coach can.
the machine behind the agents
Sign up with a wallet you already have. Connect any conformant agent. Switch in seconds. Revoke with one signature.
memory bank · consumer positioning, decision stage
Sources: memory bank 09-agent-development/reference-agents/date-coach-v2.md (live loop); this week's platform release and hiring-agent usage per the daily syncs.
07 / 11Audiences
Three audiences, one flywheel
The model on the whiteboard: creators propagate, consumers generate usage,
and the resulting numbers become the success case study that opens the doors downstream.
AConsumers
People who want an AI that works for them, not on them: clients of coaches, then the wider market. They bring usage: sessions, retention, weekly active users, the North Star.
BCreators
Coaches and domain experts with existing client books. They bring propagation: each creator onboards their clients, and keeping their brand and client book intact makes them sellers, not suppliers.
CInfra customers
Startups and enterprises who see the case study and want the same loop for their own products. They are sold downstream, by the numbers A and B generate, not by promises.
08 / 11Business model
The deal that recruits its own supply side
4
Coach verticals at launch. Cold-start is solved vertical by vertical: recruit the creators whose clients already trust them, launch their agents to those clients first.
launch scope · gtm v3
0
Zero platform capture. The brand, the method, and the client relationship stay with the coach; the platform meters usage and settles the split, nothing more.
the anti-platform deal
1 session
White-glove onboarding: from a coach's framework to a live private agent in one working session. Built with the coach, not by the coach.
the design-partner motion
The magic moment, a read that could only come from their
coach's method, drives word of mouth from the first client onward.
09 / 11Go-to-market
Work backwards from Success Case Study #1
Everything sequences toward one artifact: the first success case study, told three
ways for the three audiences. It is the singular goal that simplifies everything else.
NowDesign partners
Outreach campaigns live in two coach verticals; warm replies in play, replies drafted from the playbook, demo-first conversion.
→
Aug 8Public launch
Marketplace and vaults live. Four coach verticals at launch, magic-moment onboarding, paying demand from day one.
→
Sept#A + #B traction
The two validated numbers: consumers active weekly, creators live with their clients. Stats become tellable.
→
OctSuccess C.S. #1
One story, three angles: what consumers got (A), what creators earned (B), what the numbers prove to infra customers (C).
- North Star: weekly active users. Traction is usage, not signups.
- Content is the delivery system: every milestone above ships its story the week it happens, from the same memory bank this deck is generated from.
- Nothing is oversold: case studies use real, attributable people and live numbers, or they wait.
Sources: memory bank 07-go-to-market/launch-plan.md (dated beats, 2026-08-08); specs/gtm v3 (verticals, magic moment, weekly-active North Star); whiteboard session 2026-07-17 (A · B · C, Success C.S. #1).
10 / 11Landscape
Generic AI has scale. Platforms have supply. Nobody has both sides of trust.
| Generic assistants | Creator platforms | ManyKind.ai |
|---|
| Whose method | Nobody's: one model for everyone | The creator's, but platform-owned | The creator's, creator-owned |
| Whose data | The vendor's | The platform's | The user's: owned and revocable |
| Switching cost | Total: memory dies with the account | High: audience is captive | A signature |
| Creator economics | None | Heavy platform cut, brand submerged | Creator-first, name on the door |
| Accountability | Opaque | Opaque | Every operation signed and audited |
The position nobody else can occupy: personal AI where
the person and the expert both keep what is theirs. Generic assistants cannot offer creator
economics; creator platforms cannot offer user ownership; neither can offer both.
11 / 11The 90 days
Build more, sell more, and let the stories compound
AugLaunch Public launch Aug 8. Four verticals live, first paying demand, launch story told everywhere at once.
→
SeptTraction Hit #A and #B. Publish the real numbers weekly, wins and gaps both.
→
OctCase study Success C.S. #1 ships in three angles. It becomes the sales asset for every audience.
→
Nov–DecCompound More verticals, more creators, case study #2 and #3. The machine repeats: build, show, sell.
Decide today
Set #A and #B: the consumer and creator targets the whole plan works backwards from.
Protect the gate
The launch-critical path: coach demo for warm leads, payments to real sign-up, recruiter page to production.
Keep it in the memory bank
This deck is a measuring stick: regenerate it as the memory bank updates. If a number is wrong here, fix the memory bank, not the slide.
The first set of success stories is what really matters.
strategy session · 2026-07-17