VentureMind — AI Brain Behind Nomad Flow

Version 1.0 · May 2026 · Public

---

What Is VentureMind?

VentureMind is the autonomous AI engine powering Nomad Flow — a global platform that takes founders from raw idea to live, running business in a single session.

It does this by coordinating 10 specialist AI swarms, each responsible for one domain of building a global company:

#	Swarm	What It Does
1	Central Swarm	The executive — coordinates all other swarms
2	Engineering Swarm	Builds your product: web, mobile, infrastructure
3	Legal Swarm	Sets up your LLC, IBC, or holding structure
4	Financial Swarm	Tax strategy, bookkeeping, financial planning
5	Capital Swarm	Raises funding, manages runway and burn
6	Growth Swarm	SEO, content, paid acquisition
7	Web3 Swarm	Crypto wallets, DeFi, token strategy
8	Wealth Swarm	Personal investing, portfolio management
9	Mobility Swarm	Residency, banking, travel logistics
10	Journey Swarm	Health, lifestyle, nomad infrastructure

---

Architecture

VentureMind is built on three open-source frameworks:

Framework	Role
Paperclip	Agent org chart — defines every agent's role, reports, skills, and schedule
Feynman	Research layer — verifies claims, audits sources, reproduces data
agent-skills-eval	Validation — proves every skill actually works before shipping
gstack	Engineering team — 23 opinionated slash commands for shipping velocity

---

Repository Structure

VentureMind/
├── personas/                    ← Human-readable agent descriptions (14 agents)
│   ├── README.md                ← Swarm of Swarms index
│   ├── central-swarm-lead.md
│   ├── engineering-swarm-lead.md
│   ├── legal-swarm-lead.md
│   ├── financial-swarm-lead.md
│   ├── capital-swarm-lead.md
│   ├── growth-swarm-lead.md
│   ├── web3-swarm-lead.md
│   ├── wealth-swarm-lead.md
│   ├── mobility-swarm-lead.md
│   ├── journey-swarm-lead.md
│   ├── strategy-agent.md
│   ├── operator-agent.md
│   └── talent-agent.md
│
├── paperclip/venturemind/       ← Paperclip company package (AI org chart)
│   ├── COMPANY.md               ← Root company definition
│   ├── agents/                  ← 12 agent definitions (AGENTS.md format)
│   │   ├── central-swarm/
│   │   ├── engineering-swarm/
│   │   ├── legal-swarm/
│   │   ├── financial-swarm/
│   │   ├── capital-swarm/
│   │   ├── growth-swarm/
│   │   ├── web3-swarm/
│   │   ├── wealth-swarm/
│   │   ├── mobility-swarm/
│   │   ├── journey-swarm/
│   │   ├── watchdog/
│   │   └── orchestrator/
│   ├── teams/                   ← 6 team groupings
│   ├── projects/                ← 5 active projects
│   ├── tasks/                   ← 4 recurring scheduled tasks
│   └── skills/                  ← 4 core skills (SKILL.md format)
│
├── feynman/                     ← Feynman research agent package
│   ├── agents/                  ← 4 specialist research agents
│   │   ├── researcher/
│   │   ├── reviewer/
│   │   ├── writer/
│   │   └── verifier/
│   ├── skills/                  ← 4 feynman-powered skills
│   │   ├── venturemind-deepresearch/
│   │   ├── venturemind-lit/
│   │   ├── venturemind-watch/
│   │   └── venturemind-recipe/
│   └── research/               ← 2 active research projects
│       ├── uae-dubai-ibc-vs-wyoming-llc/
│       └── treaty-network-analysis/
│
├── gstack/                     ← gstack engineering team integration
│   ├── CLAUDE.md               ← gstack configuration for Nomad Flow
│   └── skills/                 ← 4 VentureMind-specific gstack skills
│       ├── venturemind-office-hours/
│       ├── venturemind-security-audit/
│       ├── venturemind-review/
│       └── venturemind-ceo-review/
│
├── ingestion/                   ← Structured Discovery System
│   ├── SPEC.md                 ← Full ingestion system specification
│   ├── ONBOARD-UI-SPEC.md      ← Dark terminal UI design spec
│   ├── prompts/
│   │   └── strategist-agent.md ← Strategist Agent prompt architecture
│   ├── workflows/
│   │   └── verification-loops.md ← Phase 2: 4-parallel verification loops
│   └── templates/
│       └── intake-report.md     ← Audit-ready intake report template
│
├── execution/                  ← Green Button Autonomous Execution
│   ├── README.md               ← Execution system overview
│   ├── SafetyNet-HITL.md       ← Human-in-the-Loop trigger matrix
│   ├── manifests/              ← Execution manifests
│   │   ├── EXEC-SOVEREIGN-ACCOUNT.md
│   │   ├── EXEC-ENTITY-FORMATION.md
│   │   └── EXEC-NEOBANK-FIAT.md
│   └── agents/
│       └── watchdog-agent.md   ← Real-time threat detection
│
├── security/                   ← Hardened Security Architecture
│   ├── SECURITY-VAULT.md       ← Zero-Knowledge Vault & Ghost Protocol
│   └── agents/
│       └── watchdog-agent.md   ← Watchdog Agent (separate copy)
│
├── evals/                      ← agent-skills-eval validation suite
│   ├── README.md
│   ├── safety-net-protocols/   ← SafetyNet HITL skill eval
│   ├── orchestrator/           ← Orchestrator skill eval
│   ├── green-button/           ← Green Button execution eval
│   └── vault-access/          ← Zero-Knowledge Vault eval
│
├── SPEC.md                    ← Master technical specification
└── PRD.md                    ← Product Requirements Document (PDF attached)

---

The 3 Modes: Advise → Authorise → Execute

Mode 1: Advise (Discovery)

The Strategist Agent runs a 7-stage intake interview. Every answer is verified against 4 primary sources in parallel. An audit-ready intake report is generated with a confidence score.

Mode 2: Authorise (SafetyNet)

The Green Button system presents every proposed action with full context, risk level, and legal implications. The user approves or rejects each action. CRIMSON-level actions trigger mandatory human review.

Mode 3: Execute (Autonomous)

Approved actions run autonomously. Every step is logged to an immutable decision chain. The Watchdog Agent monitors every action in real time and can pause any execution instantly.

---

Core Skills

Skill	What It Does
safety-net-protocols	Enforces HITL triggers, confidence thresholds, and escalation paths
orchestrator	Routes tasks to the correct swarm, manages inter-swarm dependencies
green-button	Manages the full execution manifest lifecycle
vault-access	Zero-knowledge document handling with ghost redaction

---

The 10 Pillar Services (Nomad Flow)

1. Idea → Launch — Product development with engineering swarm
2. Legal Counsel — Entity formation, contracts, compliance
3. Nomad Navigator — Visa, residency, travel logistics
4. Luxury Optimizer — Lifestyle infrastructure (flights, hotels, concierge)
5. Health — Health insurance, telemedicine, fitness
6. Tax Optimisation — International tax structure, treaty optimisation
7. Crypto Services — Wallets, exchanges, DeFi, tax reporting
8. Entity Formation — LLC, IBC, holding structures globally
9. Capital Raise — Investor matching, pitch deck, due diligence
10. The Accountant — Bookkeeping, CFO services, financial reporting

---

Build Roadmap

Phase	Focus	Status
1	Core Agent Brain — intake, routing, orchestration	Complete
2	SafetyNet — HITL triggers, compliance layer	Complete
3	Green Button — execution manifests, real execution	Complete
4	Full Platform — UI, UX, payment, onboarding	In Progress

---

Stack

Backend: Python · FastAPI · LangGraph · PostgreSQL · Redis · Qdrant
Frontend: React · TypeScript · Tailwind CSS
AI: Claude · Groq · OpenAI
Agents: Paperclip · Feynman · gstack
Infrastructure: Docker · GCP · Cloudflare

---

Revenue Model

Tier	Price	Includes
Nomad	$99/mo	1 swarm, basic intake, 10 actions/mo
Founder	$299/mo	3 swarms, full intake, 50 actions/mo, SafetyNet
Sovereign	$999/mo	All 10 swarms, unlimited actions, Watchdog, priority support

Plus: % of entity formation fees, % of capital raised, affiliate commissions.

---

Related Documents

• SPEC.md — Full technical architecture
• PRD.md — Product Requirements Document (plain English)
• ingestion/SPEC.md — Structured Discovery System
• execution/SafetyNet-HITL.md — Human-in-the-Loop framework
• security/SECURITY-VAULT.md — Zero-Knowledge security architecture

---

VentureMind is the AI brain behind Nomad Flow. Built for founders who want to build globally, not just locally.