Healthcare Delivery for AI Agents

Arlo is the clinical action API for AI health platforms. It gives your AI its own clinicians, like a hospital does for patients.

from arlo import Arlo

client = Arlo()

encounter = client.encounters.create(
    patient_state="CO",
    type="prescription",
    condition="strep_throat"
)

It's not AI for your clinic.
It's a clinic for your AI.

Your platform triages, educates, and recommends. Arlo handles the part that requires a license.

Encounters API

Create, manage, and resolve clinical encounters entirely via API.

  • Prescriptions
  • Lab orders
  • Referrals
  • Consultations
  • Renewals
  • Follow-ups
  • Webhooks
  • SDKs + MCP

Connect through MCP

Add the Arlo MCP server to your agent and it gets native access to create encounters, check status, and receive results. No SDK required.

{
  "mcpServers": {
    "arlo": {
      "type": "command",
      "command": "npx -y arlo-mcp",
      "env": {
        "ARLO_API_KEY": "arlo_sk_xxxxxxxxx"
      }
    }
  }
}
{
  "mcpServers": {
    "arlo": {
      "command": "npx",
      "args": ["-y", "arlo-mcp"],
      "env": {
        "ARLO_API_KEY": "arlo_sk_xxxxxxxxx"
      }
    }
  }
}
from arlo import Arlo

client = Arlo(
    api_key="arlo_sk_xxxxxxxxx"
)

encounters = client.encounters.list(
    status="pending"
)

Three API calls. Real clinical action.

1

Create encounter

Your AI identifies a clinical need. POST to /encounters with patient state, condition, and encounter type.

2

Arlo routes

We match to a licensed clinician in the patient's state. Credentialed, malpractice-covered, and ready.

3

Clinician acts

Prescription sent. Lab ordered. Referral placed. You get a webhook with the result.

Request
POST /v1/encounters

{
  "patient_state": "CO",
  "type": "prescription",
  "condition": "strep_throat",
  "context": {
    "symptoms": ["sore_throat", "fever"],
    "duration_days": 3,
    "severity": "moderate"
  }
}
Response
{
  "id": "enc_a1b2c3d4",
  "status": "completed",
  "type": "prescription",
  "clinician": {
    "type": "NP",
    "state_license": "CO"
  },
  "actions": [
    {
      "type": "prescription",
      "medication": "Amoxicillin 500mg",
      "pharmacy": "CVS #4821",
      "status": "sent"
    }
  ],
  "completed_at": "2026-02-26T14:32:00Z"
}

Built for healthcare at scale.

50

States covered through our PC network.

<2 hrs

Median encounter resolution time.

100%

Clinician review on every encounter. No exceptions.

0

Lines of healthcare compliance code you write.

Every AI health platform hits the same wall.

Your AI can assess symptoms, coach patients through chronic conditions, and generate care plans. Then someone needs a prescription. Or a lab order. Or a referral.

That's where your product breaks.

Not because your AI isn't good enough. Because healthcare requires a licensed human to act. Every state has different rules. Every action needs credentials, malpractice coverage, and compliance infrastructure you don't have.

Arlo is that infrastructure. One API. Licensed clinicians in all 50 states. Prescriptions, labs, and referrals — completed and returned to your platform via webhook.

You build the AI. We handle the license.

We handle the hard part.

HIPAA compliant

End-to-end encryption, BAAs, audit trails. Already certified.

50-state CPOM coverage

Corporate Practice of Medicine compliance in every state through our PC network. You don't need your own medical corporation.

AI transparency

Colorado SB 24-205, informed consent, AI audit trails. Documented and maintained.

FDA CDS exempt

Clinical Decision Support exemption under 21st Century Cures Act. Documented and on file.

Malpractice covered

Every clinician carries malpractice insurance. AI-assisted workflows included.

State telehealth law

We track registration requirements across every state so you don't have to.

From symptom checkers to autonomous agents.

Symptom checkers

Your AI assesses, then hands off to Arlo when a prescription, lab, or referral is needed. The patient never leaves your app.

Health coaching platforms

Chronic condition management that can actually act — refill a prescription, order bloodwork, adjust a care plan.

Pediatric apps

Parents get answers from your AI and treatment from our clinicians. Text-based, no waiting rooms.

AI agents

Your agent identifies a clinical need, calls our API, and resolves it. No human in your loop. Licensed human in ours.

Pay per encounter.

Build

For startups validating clinical workflows.

  • First 50 encounters free
  • Then per-encounter pricing
  • Community support
Get Started

Enterprise

For organizations with complex requirements.

  • Custom contracts
  • White-label options
  • Dedicated clinician pools
  • On-call clinical team
Contact Sales

Frequently Asked Questions

Do I need my own medical corporation to use Arlo?

No. Arlo operates through its own Professional Corporation (PC) network, which provides Corporate Practice of Medicine (CPOM) compliance in all 50 states. You integrate via API — we handle the legal entity structure.

How does malpractice coverage work?

Every clinician in the Arlo network carries their own malpractice insurance that covers AI-assisted clinical workflows. You don't need to procure separate coverage for encounters processed through our API.

What states are supported?

All 50 US states plus Washington, D.C. Our PC network maintains active licenses and telehealth registrations in every jurisdiction. When you POST an encounter with a patient state, we route to a clinician licensed there.

How long does an encounter take to resolve?

Median resolution time is under 2 hours. Urgent encounters are prioritized. You receive a webhook when the clinician completes their review and takes action.

Is Arlo HIPAA compliant?

Yes. End-to-end encryption, signed BAAs, full audit trails, and SOC 2 compliance. We execute a Business Associate Agreement with every customer before processing PHI.

Can I use Arlo with my existing AI agent framework?

Yes. Arlo works via REST API, Python SDK, or MCP server. If your agent can make HTTP requests or connect to an MCP server, it can use Arlo. We support Cursor, Claude Code, and any MCP-compatible tool.

What happens if a clinician declines an encounter?

If a clinician determines the encounter requires in-person care or falls outside telehealth scope, the encounter is returned with a status of declined and a reason code. You're not charged for declined encounters.

Give your AI a medical license.

One API to prescribe, order labs, and refer patients. In all 50 states. With full compliance.