Ship patient intake in minutes with an LLM-powered API
Replace $35–45K/yr VA workflows with an intake pipeline that processes unstructured forms into structured EHR-ready JSON
$npm install @holistic/intake-sdk HIPAA-aligned architecture from day one
Handles OCR extraction, categorization, and structured output. Display-only architecture limits PHI exposure. Requires Business Associate Agreement with your hosting provider.
- PHI-minimal design
- BAA-compatible flow
- Audit trail included
```bash
curl -X POST https://api.holistic.example/v1/intake \
-H 'Authorization: Bearer sk_live_...' \
-F 'file=@patient_form.pdf'
```Structured extraction without fine-tuning
Leverages GPT-class models with a practitioner-knowledge prompt layer. Outputs symptoms, diagnoses, and intake flags as normalized categories. Handles hand-written and scanned intake forms.
- GPT-4o compatible
- Category normalization
- Confidence scores on output
```typescript
const result = await intake.process({
pdfUrl: 'https://storage.example/inake.pdf',
extractionSchema: ['symptoms', 'medications', 'history']
});
console.log(result.data);
// { symptoms: [...], medications: [...], history: [...] }
```Build your intake flow in <50 lines
POST a form URL or file. GET back patient data as structured JSON. No frontend lock-in. Webhooks push results to your existing practice management system.
- REST + Webhooks
- JSON in/out
- Works with Jane App, Mindbody, PracticeMuse
```typescript
// Create a patient intake session
const session = await intake.sessions.create({
practitionerId: 'pr_abc123',
formTemplate: 'default-holistic',
patientEmail: '[email protected]'
});
// Webhook fires when intake is complete
// POST /webhooks/intake.complete → structured patient record
```Implementation FAQ
What HIPAA obligations apply if I use this?
Because the API processes display-only (no storage of PHI), you may qualify for reduced obligation scope. However, you remain the covered entity. A Business Associate Agreement with your hosting provider is required. Legal counsel should review your specific architecture before launch.
What does an intake call cost at scale?
At current GPT-4o pricing, a standard intake form (2–3 page PDF) costs approximately $0.01–0.02 per call. At 100 intakes/month, that is ~$1–2/month in API costs. Volume pricing tiers are available for practices exceeding 500 intakes/month.
Does this replace my practice management software?
No. This intake pipeline feeds into existing practice management systems via webhook or API import. It solves the front-end intake gap without requiring a migration away from Jane App, Mindbody, or similar platforms your clients already use.
What form types does the LLM layer support?
Current extraction models handle standard structured intake forms (PDF, scanned documents, embedded images). Hand-written responses with legible text are supported with ~92% accuracy. Poorly scanned or non-English forms may require a custom fine-tuning engagement.
Ship your intake pipeline this month
Open-source SDK on npm. No vendor lock-in. Deploy to any Node.js environment.