Pre-Seed Round Open

Trust Infrastructure for Regulated Healthcare AI

$7M at $40M Post-Money

With paid Premium pilot signed: $8M at $45M post-money

Pre-action enforcement. Post-inference verification. Cryptographic audit. Controlled production live. GMAI is the middleware that governs patient-facing AI before and after the model call.

Schedule Investor Meeting
Patent Filed
Controlled Production Live
Soft Launch Aug 2026
EU AI Act: 27 days
The Problem

AI Is Powerful. But Not Trustworthy.

Individual Crisis

  • 1.1B people need mental health support
  • $260-400/month for therapy
  • Weeks-long waitlists for care

Enterprise Crisis

  • Zero AI governance leaders in market
  • No cryptographic compliance solutions
  • Massive liability exposure without audit trails

Regulatory Hammer

  • EU AI Act high-risk enforcement
  • Up to €30M fines for non-compliance
  • Healthcare AI must meet strict standards
27 days until enforcement
$238B

The Health Literacy gap costs the U.S. economy up to $238B annually. Only 12% of adults are proficient. Nobody owns this category.

Source: CHCS, 2024 / NAAL

Trust is promised, not proven. Privacy is a feature, not architecture. Every enterprise deploying AI in regulated industries needs a trust layer.

The Platform

One Trust Layer. Every Regulated AI Application.

The Middleware

GMAI

MiAngel Middleware AI

The patent-pending cryptographic control plane. Every inference is intercepted before it reaches the model (pre-action enforcement), every output validated against patient context and policy before delivery (post-inference verification), and every interaction written to a SHA-256 hash-linked, RSA-PSS signed audit chain.

The Applications

DeBrah Personal + Clinical

The First Applications

DeBrah Personal is the consumer reference implementation: self-pay, controlled production, public soft launch August 2026. DeBrah Clinical brings the same governed layer to physician practices, payers, and care management. The architecture extends to every regulated AI vertical.

GMAI is the middleware platform. DeBrah Personal proves it, DeBrah Clinical scales it, and the same layer licenses to any regulated AI vendor. One middleware, multiple applications.

What HTTPS did for internet commerce, GMAI aims to do for regulated AI. Just as HTTPS encrypts web traffic, GMAI cryptographically governs AI interactions, before and after the model.

1

Authenticate

Who is making this request?

2

Check Consent

What data can the AI access?

3

Select Context

Salience scoring from memory vault

4

Apply Policies

What can the AI say and do?

5

Log Everything

Audit trail for compliance

6

Assess Risk

Graded safeguards, escalation if thresholds breached

Only AFTER GMAI validates does the prompt reach the AI model. No other system provides this pre-model governance layer.

Live in Production

This Is Not a Mockup.

300K+
Lines of TypeScript
Production code, not slides
<100ms
API Response
Real-time trust pipeline
8
Claimed Mechanisms
In controlled production
14
Pipeline Steps
Full GMAI enforcement chain
AI Companion Chat

AI Companion Chat

Mood Tracking

Mood Tracking

Encrypted Journal

Encrypted Journal

Wellness Dashboard

Wellness Dashboard

The Moat

Three Stacked Moats. Each One Compounds.

Layer 1: Patent-Pending IP

The legal moat. U.S. Patent Application #19/385,439, Track One expedited. Eight claimed mechanisms implemented in controlled production; continuation filed. Claims directed to pre-model cryptographic enforcement as a category (scope subject to examination). Every month in production strengthens the execution evidence.

Layer 2: Architecture as IP

The technical moat. Hexagonal microservices middleware, pre-model enforcement engine, cryptographic audit chain (SHA-256, RSA-PSS), attestation-gated memory vault. Replicating it means rebuilding the inference pipeline around pre-model enforcement, with a 24+ month head start on hexagonal extraction.

Layer 3: Clinical Distribution

The operator moat, the one capital cannot quickly buy. A co-founder CMO with 6 medical centers, a 14-site research network, and a B2B platform reaching ~20,000 providers. Day-one deployment access through co-founder alignment (arm's-length, board recusal). Every deployment adds switching costs.

Patent Application #19/385,439 | Track One

Attestation-Gated Memory

Conversation history cryptographically bound to identity. Cannot be accessed without biometric + device proof.

Control Segments

Canonical signed headers attached to every LLM request. Contains: auth state, consent scope, policy parameters.

Salience Scoring

Deterministic algorithm for context selection. Auditable memory recall, not a black box.

Audit Chain

Every interaction generates a tamper-evident log. Cryptographically signed + timestamped.

Pre-Model Enforcement

Trust layer operates BEFORE the LLM receives the prompt. Model-agnostic. Works with any AI provider.

Dr. Luis O. Noguerol

"GMAI's pre-model enforcement architecture represents the most rigorous cryptographic governance framework I have evaluated in 25 years of federal cybersecurity. This is not incremental improvement. This is an architectural shift in how we trust AI."

Dr. Luis O. Noguerol

C|CISO of the Year, EC-Council. 100+ Cybersecurity Certifications.

The Opportunity

$187.7B Healthcare AI Market

$187.7B
Total Addressable by 2030
AI in Healthcare, 38.5% CAGR (Grand View Research 2025)
$35B
Serviceable Addressable by 2030
US RPM $29B + Healthcare AI Governance $6B (MarketsandMarkets + Gartner)
500K+
Initial Reachable Surface, Y1-Y3
20K+ providers, 60 active trials (co-founder channels; arm's-length, board recusal)
TAM > SAM > Reachable Surface. Sourced sizing, not bottoms-up guesses.
The horizontal infrastructure category is still undefined. MiAngel is building it.

MiAngel + GMAI

  • Pre-action cryptographic governance: runtime enforcement before the model
  • Clinical distribution: co-founder healthcare network, day-one access
  • Horizontal platform: extensible to N regulated-AI applications
  • Patent-pending architecture in controlled production

Healthcare AI Products

Hippocratic AI, Ambience, Suki, Abridge

  • Vertical products, not infrastructure
  • Each rebuilds the same compliance layer
  • No pre-action cryptographic enforcement
  • None can sell the layer

Governance & AI Gateways

OneTrust, Drata, Credo AI, Cloudflare AI

  • Documentation and policy, not runtime enforcement
  • Cannot stop an inference before it happens
  • No clinical distribution access
  • No patient context

A new category intersection, with no defined incumbent.

Three Revenue Streams

One Platform. Three Revenue Lines.

DeBrah Personal

Consumer Subscription

$19.99/mo Standard, $34.99/mo Premium+

Gross Margin80-90%
BillingStripe direct
Modeled Y3~60K subscribers

Public proof point. Hispanic / Cuban-American beachhead. Modeled, subject to conversion and retention.

DeBrah Clinical

Reimbursement-Aligned (modeled)

$30-60 PMPM gross/patient (modeled)

BasisCMS RPM/RTM-aligned
CPT99453-58, 98975-81
StatusUnder validation

Primary near-term revenue (modeled). Gross opportunity does not equal MiAngel net revenue; subject to pilot validation, payer policy, and commercial agreement.

GMAI Platform

Enterprise Licensing

$50K-$500K modeled enterprise ACV

Federal pilots$250K-$2M
Gross Margin90%+
ModelPer-seat or rev share

Long-term platform revenue line. Highest LTV per customer. Category-defining moat.

Three revenue streams, one platform. CMS RPM/RTM-aligned clinical revenue is modeled and under validation; platform licensing is the long-term margin engine. Reimbursable economics apply only if validated.

The Deployment Surface

Three Deployment Lanes. One Operating Surface.

Three revenue mechanisms, each already inside the deployment surface our CMO operates. Premium Healthcare, IronKloud, and BioResearch Partner are Hoover-controlled related parties; commercial deployments operate at arm's-length with board recusal.

Clinical Deployment

Premium Healthcare

ProductDeBrah Clinical, patient-facing AI governed by GMAI
Deployment Surface500K+ patients, 6 medical centers in Miami-Dade, 300+ employees
GTM StatusMSA in execution. Paid pilot target: Q3 2026.
B2B Integration

IronKloud Platform

ProductGMAI Platform, middleware licensed to AI features inside IronKloud
Deployment Surface~20,000 healthcare providers (medical, dental, veterinary)
GTM StatusPipeline: discussions to LOI (not committed revenue). Integration target: 2027.
Clinical Research

BioResearch Partner

ProductDeBrah as conversational PRO (patient-reported outcomes)
Deployment Surface60 active clinical trials, 90 in queue, 14-site research network
GTM StatusPilot trials achievable in 6 months (modeled, subject to sponsor/CRO validation).

Access is not revenue until contracts, compliance workflows, and billing mechanics are validated.

What We Have Cleared

Most Pitches Show Roadmap. We Show Shipped.

Controlled production live at miangel.ai
Eight claimed mechanisms implemented in controlled production code
GMAI standalone service, staging-live
Stripe billing live, consumer subscription flow operational
U.S. Patent Application #19/385,439 filed (Track One expedited)
HIPAA BAAs executed: OpenAI and Google Vertex AI (Anthropic in execution)
MIANGEL trademark, Notice of Allowance issued
SAM.gov active (CAGE 19RX3); SBA 8(a) submitted
Co-founder triangle formed (CEO, CTO, CMO)
Premium Healthcare MSA (in execution)
Public soft launch scheduled: August 11, 2026
Aug 2026

Public Soft Launch

DeBrah Personal public launch, August 11, 2026

M6

Early Deployment

Premium MSA paid pilot signed, DeBrah Clinical at first 2 medical centers, 1,000+ active patients, SOC 2 Type II kicked off

M12

Scale Inflection

10,000+ active clinical patients, IronKloud B2B integration LOI, first BRP research pilot, ~$3M ARR run rate

M18

Series A Trigger

~$5-7M ARR run rate, Premium full deployment, IronKloud B2B revenue, Series A roadshow at $200M+ target

Completed items are verifiable today. In-execution and scheduled items are labeled as such and documented in the data room.

In the News

In the News

Press Release · June 2026

Keila Hoover, MD, Joins MiAngel as Co-Founder and Chief Medical Officer

Distributed via ACCESS Newswire, syndicated across
The Team

Built by Operators

Rafael Mas, Founder and CEO

Rafael Mas

Founder & CEO

  • 25+ years enterprise infrastructure (Royal Caribbean, Citrix, Intermex)
  • Built 300K+ lines of production TypeScript solo
  • HIPAA/SOC-2/GDPR/NIST domain expert
  • Patient Zero: uses DeBrah daily for anxiety management
Read the full founder story
Keila Hoover, MD, Co-Founder and CMO

Keila Hoover, MD

Co-Founder & CMO

  • Board-Certified Family Medicine, 23 years clinical, Clinical Assistant Professor at FIU College of Medicine
  • Operates Premium Healthcare: 6 medical centers in Miami-Dade, 500K+ patients
  • BioResearch Partner: 14-site clinical research network, 60 active trials
  • IronKloud co-founder: B2B platform reaching ~20,000 healthcare providers
Nikhil Misra, Co-Founder and CTO

Nikhil Misra

Co-Founder & CTO

  • 20+ years enterprise architecture (Oracle, J&J, Pepkor Group)
  • Founded Creative Flakes, $2.7M USD client revenue
  • Architected AI systems at 95% accuracy (Vertex AI, Gemini, GCP)
  • Led mission-critical integration processing $500M+ annually
Advisor
Dr. Luis O. Noguerol, Security and Governance Chief Advisor

Dr. Luis O. Noguerol

Security & Governance Chief Advisor

  • C|CISO of the Year worldwide, 2016
  • 100+ cybersecurity certifications
  • ISSO at NOAA (active federal), FedRAMP expert
  • Independent technical validation: NIST, FISMA, HIPAA mapping
5.0Google Reviews from early testers
The Ask

$7M Pre-Seed

40%$2.8M
Engineering
3-4 engineers, GMAI infrastructure, production deployment
15%$1.05M
Clinical & Compliance
SOC 2 Type II, HITRUST CSF prep, clinical advisory board
15%$1.05M
Healthcare GTM
Premium MSA, IronKloud integration, BRP pilot trials
10%$700K
Marketing & Brand
Brand, content, conferences, ambassador, IR
10%$700K
Operations
Legal, finance, regulatory, BAA renewals
10%$700K
Founder / Operator Comp
Market-conservative CEO, CTO, CMO salaries

Why Now

Patent-protected architecture with 18-24 month head start
EU AI Act high-risk enforcement: August 2, 2026
300K+ lines of TypeScript built and deployed
Founder-market fit: 25 years enterprise + lived mental health experience
Dual-engine model with high-margin B2B upside
Category creation opportunity: Pre-Model AI Governance

The window to own the Trust Layer for AI is open right now. In 18 months, it won't be.

The next decade of AI will be built on trust. We are building the Trust Layer.