TheoVex
01 · What is TheoVex?

Intelligence, as a layer.

theovex is an AI research lab building HIPAA-native intelligence for healthcare and insurance. One orchestration engine, four products, and the model platform all of them run on, powered by theo.

The suite

One suite.
One intelligence layer.

Four products, and the model platform all four of them run on.

Swipe · tap a card to dive in
How does it work?

One call runs the whole pipeline.

Every request moves through the same five stages: classify the intent, load the right skills, route across 300+ models, run the agent loop, and respond with citations and usage. No providers to wire, no framework to babysit.

  1. 01

    Classify

    Detect the intent and task type behind the prompt.

  2. 02

    Load Skills

    Inject domain expertise, tools, and prompt extensions.

  3. 03

    Route

    Select the optimal engine, with automatic failover.

  4. 04

    Agent Loop

    Think, act, and observe with tools until the task is done.

  5. 05

    Respond

    Return content with citations, artifacts, and usage.

Six engines, one router
Theo Flashfast

Low-latency completions for quick, high-volume tasks.

Theo Reasonthink

Deep, multi-step reasoning and analysis.

Theo Codecode

Production-quality code with an extended output budget.

Theo Createimage

Photorealistic image generation with auto-fallback.

Theo Motionvideo

Generative video, run as an async job.

Theo Researchresearch

Multi-step web research with synthesized citations.

theo.ts
import { Theo } from "@hitheo/sdk";

const theo = new Theo({ apiKey: process.env.THEO_API_KEY! });

// One call. Theo classifies, routes, and runs the pipeline.
const res = await theo.complete({
  prompt: "Summarize this patient's recent labs",
  skills: ["clinical-summary"],
  mode: "auto",
});

console.log(res.model.label);      // "Theo Reason"
console.log(res.usage.cost_cents); // 0.15

Embed it as your own with E.V.I.

Ship a branded assistant. Your users see your product, not ours.

The engine, by the numbers

Every product in the suite calls the same engine, so Theo is the only dependency any of them has.

Models orchestrated
300+Models orchestrated
API call, full pipeline
1API call, full pipeline
Stage orchestration
5Stage orchestration
HIPAA-native
100%HIPAA-native
In production

Trusted across healthcare's hardest workflows.

Carriers, agencies, clinics, and the networks between them run the suite every day. A few of the teams, and the work they ship on it.

  • Community Health ChoiceCommunity health plan
  • New Health PartnersProvider network
  • New Health MedicalMedical group
  • Hispana de SegurosInsurance agency
  • Innovate SolutionsHealthcare solutions
  • Key HealthHealth services
The work that ships on us

Health plans & carriers

Governed workspaces where plan teams draft, review, and version the work compliance has to sign.

  • Prior-auth summaries
  • Payer review decks
  • Audit-ready history

Insurance agencies

The day-to-day pipeline run through Theo CRM: quoting, enrollment paperwork, and commissions in one system.

  • Quoting & e-app
  • Commission tracking
  • Pipeline & dialer

Clinics & medical groups

Practice operations with an AI scribe on every call and scheduling that stays inside the compliance posture.

  • AI scribe notes
  • Scheduling & billing
  • Practice management

Community health programs

Outreach teams that assemble program materials same-day from locked, on-brand templates.

  • Program docs
  • Event kits
  • Outreach flyers
Teams shown run OpenCharts, the suite's workspace productopencharts.com/enterprise
What do we do differently?

We build compliance from the model layer up.

Incumbent AI tools can't enter these markets without rebuilding from the model layer up, and that gap is our moat.

The thesis

We're an AI research lab, not a wrapper. theo is an orchestrator that outruns reaching for Perplexity, and a framework that replaces LangChain. We build intelligence as a layer: HIPAA-native models, pre-tailored for the only two sectors we serve, healthcare and insurance.

01

Orchestrator, not aggregator

We compete with Perplexity by routing every task across 300+ models, picking the best engine per prompt instead of one model for everything.

02

The framework, replaced

A managed pipeline that does the work of LangChain for you: classification, routing, tools, and agent loops, behind one call.

03

Compliant by construction

HIPAA, SOC 2, and HITRUST are native to the stack, so the intelligence is safe for PHI from the first request.

A wedge competitors can’t copy

Healthcare + insurance focus

We don't build general-purpose AI. Every product is purpose-built for the highest-stakes regulated verticals, where being right and compliant is the whole game.

HIPAA + SOC 2-native moat

Compliance is built into the framework from the model layer up, not bolted on. Incumbent AI tools can't enter these markets without rebuilding, and that gap is our moat.

One intelligence layer, every product

Theo powers every product we ship, so each one compounds the same core orchestration, safety, and compliance work instead of starting over.

Inherited, not rebuilt
  • SOC 2
  • HIPAA
  • HITRUST
  • PHI-native
Still have questions?

Ask theo about
what we do.

Anything the page above didn't cover: the products, the engine, the research, or the team.

Ask Theo about
  • What we do
  • How it works
  • Our research
  • The team
  • Investors
TheoVexStart building

Build on theo.

One API key. The whole pipeline. HIPAA-native from the first request, built for healthcare and insurance.