Quickstart — Intake Agent

The Intake Agent is the Medera Patient Intake Agent — voice-first, adaptive, orchestrated via Medera Voice with Medera STT nova-2-medical transcription and an Medera voice. She follows a documented 7-state conversation state machine and registers 4 function calls.

Conversation state machine

Configure the deployment

Use the Agent Builder Canvas to clone the the Intake Agent — Patient Intake Agent template. Configure:| Field | Default | | :--- | :--- | | intake_template_id | Practice intake template | | screeners | ["phq-9", "gad-7", "audit-c", "c-ssrs"] | | routing_matrix | Practice routing rules by acuity and payer | | crisis_escalation_tree_id | On-call escalation tree | | workflow_modules | ["general_intake"] | | language | en-US | Publish the workflow — the canvas auto-deploys to sandbox at https://sandbox.medera.info/{org}/intake-v{ver}. |

Initialize the agent for the organization

bash cURL curl -X POST https://api.medera.info/api/onboarding/initialize-organization \ -H "X-API-Key: $MEDERA_API_KEY" \ -d '{ "org_name": "Acme Clinic", "org_type": "behavioral_health", "initial_agents": ["intake"], "voice_settings": { "language": "en-US", "tone": "warm" } }'

typescript JavaScript await client.onboarding.initializeOrganization({ org_name: 'Acme Clinic', org_type: 'behavioral_health', initial_agents: ['intake'], voice_settings: { language: 'en-US', tone: 'warm' }, });

python Python client.onboarding.initialize_organization(org_name="Acme Clinic", org_type="behavioral_health", initial_agents=["intake"], voice_settings={"language": "en-US", "tone": "warm"},)

Start an intake call

bash cURL curl -X POST https://api.medera.info/api/intake-calls \ -H "X-API-Key: $MEDERA_API_KEY" \ -d '{ "patient_id": "pat_abc", "phone_number": "+15555550100", "deployment_id": "dep_abc" }'

typescript JavaScript const call = await client.intake.createCall({ patient_id: 'pat_abc', phone_number: '+15555550100', deployment_id: 'dep_abc', });

python Python call = client.intake.create_call(patient_id="pat_abc", phone_number="+15555550100", deployment_id="dep_abc",)

Subscribe to real-time updates

const ws = new WebSocket(`wss://api.medera.info/ws/call/${call.id}?token=${apiKey}`);
ws.addEventListener('message', (e) => {
  const event = JSON.parse(e.data);
  // event.type: 'call-started' | 'utterance-processed' | 'task-extracted' | 'transcript-update' | 'call-ended'
});

Receive outbound webhooks

If your tenant has a webhook URL configured, you’ll receive signed events:

http POST <your_webhook_url> Medera-Signature: sha256=<hmac> Medera-Event: intake.summary.ready Content-Type: application/json { "event": "intake.summary.ready", "call_id": "call_abc", "summary": { "screeners": { "phq-9": { "total": 14, "severity": "moderate" } }, "icd10_candidates": ["F32.1"] } }

See Medera Voice webhooks.

The Intake Agent’s 4 function calls

Function	Parameters	Purpose
`verify_identity`	`patient_id`, `verification_data`	2-of-3 ID confirmation (DOB, address, SSN last 4)
`document_symptom`	`symptom_description`, `severity_level`, `onset_date`	Capture chief complaint
`flag_for_review`	`reason`, `urgency_level`	Flag for clinical review queue
`escalate_urgent`	`escalation_reason`, `call_action`	Hard escalation for emergencies (no agent gating)

Compliance guarantees

Verbal recording + AI disclosure is mandatory before any clinical questions. Identity verification requires 2-of-3 confirmation. Crisis-language detection bypasses agent gating and triggers escalate_urgent immediately.

What’s next

Patient Intake Agent

Agent-level documentation.

Medera Voice Webhooks

Webhook event schema.

Intake API

REST surface.

Crisis Response Agent

Crisis-language detection.

​Conversation state machine

​The Intake Agent’s 4 function calls

​Compliance guarantees

​What’s next

Patient Intake Agent

Medera Voice Webhooks

Intake API

Crisis Response Agent

Conversation state machine

The Intake Agent’s 4 function calls

Compliance guarantees

What’s next