Technical Documentation

AI-powered data integration
built around Healthie EHR

A HIPAA-compliant, AI-powered ETL engine that reads from and writes to Healthie and dozens of third-party platforms — no engineering team required on your side.

Healthie EHR
source / target
API key
⚡ AI Integration Engine
Google Cloud · BAA covered
API key
3rd Party Platform
HubSpot, Stripe, BigQuery…
Instructions
text / markdown

Setup in five steps

There is no self-service UI at this stage — our team handles the configuration end-to-end. Here is precisely what happens, in order.

01

You provide a Healthie API key

We require a Healthie API key scoped to the data your integration needs to read or write. Our team will guide you on the minimum required permissions. The key is stored encrypted at rest inside our Google Cloud infrastructure.

↳ secure handoff ↳ assisted by our team
02

You provide a credential for the third-party system

If the integration targets an external platform (e.g. HubSpot, Stripe, BigQuery), you provide the relevant API key or OAuth credential. If both source and destination are within Healthie itself, this step is skipped.

↳ secure handoff
03

You describe your business logic in plain text or Markdown

Write a document — in natural language, structured Markdown, or bullet points — describing how each Healthie entry (a patient record and all associated data, excluding session notes) should be processed. This covers field mappings, transformation rules, conditional logic, custom attributes, and computed values.

There is no schema to learn, no configuration DSL to master. If you can describe what you want in writing, we can integrate it.

↳ reviewed & refined by our team
04

Our team converts your instructions into a precise AI prompt

We work with you to turn your business logic document into a crisp, unambiguous instruction set that the AI agent will follow on every run. Ambiguities are resolved before the integration goes live. This is a collaborative step — we iterate until the behaviour matches your expectations exactly.

↳ collaboration required ↳ one-time setup
05

The integration runs on a configurable schedule

On each scheduled run, the engine deploys an AI agent that iterates over every Healthie entry that has changed since the previous run. For each changed entry the agent applies your instructions and writes the result to the target system. Unchanged records are skipped.

Schedules range from hourly to weekly. Monthly schedules are not currently supported. Batch frequency is adjustable at any time by contacting our team.

↳ fully automated after setup

What is in a Healthie entry

Each processing unit is a patient record plus the full set of data associated with that patient — with one exclusion.

Included in processing

  • Patient demographics and profile fields
  • Custom attributes and metadata
  • Insurance and billing information
  • Appointments and scheduling data
  • Care team assignments
  • Tags, labels, and flags
  • Any other structured patient data
🔒

Excluded from processing

  • Session notes and clinical documentation

Session notes are excluded to minimise PHI surface area and to focus the integration on operational data flows rather than clinical records.

Google Cloud, HIPAA-covered

The entire integration platform runs on Google Cloud. We operate under a Business Associate Agreement (BAA) covering the handling of Protected Health Information as required under HIPAA.

Google Cloud Platform

All compute, storage, and orchestration

BAA in place

HIPAA Business Associate Agreement covers all PHI handling

Credentials encrypted at rest

API keys stored with envelope encryption

No UI (currently)

All config managed by our team — UI on the roadmap

ℹ️

Roadmap note: A self-service management UI is planned for a future release. Currently all configuration — including credential updates, schedule changes, and instruction revisions — is handled by our team. Changes can be submitted at any time and are processed with no additional lead time.

Platforms and use cases

The engine ships with pre-built connectors for the platforms below. Custom connectors can be added on request.

🤝

CRM — HubSpot

Bi-directional sync. Create or update HubSpot contacts when Healthie patient profiles change, and keep Healthie in sync as HubSpot attributes are updated. Full control over field mapping and custom attribute computation.

📊

Big data & analytics

Push structured patient data to Google Cloud BigQuery, AWS S3, AWS Redshift, SQL databases, or MongoDB. Add-ons cover CSV, Google Sheets, Excel, and Airtable.

💬

Messaging & marketing

Export contact details to Brevo, HubSpot, customer.io, and similar platforms to power communication workflows triggered by patient events in Healthie.

💳

Billing platforms

Export patient and payer data into Stripe, Candid Health, Nirvana, and STEDI for claims processing, revenue cycle management, and direct billing.

A sample of connected platforms:

Healthie EHR HubSpot Kustomer Stripe Candid Health Nirvana STEDI BigQuery AWS S3 AWS Redshift MongoDB PostgreSQL Brevo customer.io Google Sheets Airtable + custom

Batch run frequency

Integrations run in batch mode. Each run processes only records that have changed since the previous run. The following frequencies are available.

Frequency Available Typical use case Notes
Hourly Near-real-time CRM or billing sync Higher volume cost tier
Every 6 hours Intraday analytics refresh
Daily Overnight reporting, end-of-day billing exports Most common schedule
Weekly Low-cadence data warehousing
Monthly Not currently supported
⚠️

Monthly schedules are not supported. The minimum granularity is weekly. If your use case requires monthly cadence, contact us — a workaround using a weekly run with suppressed output on non-month-end weeks may be feasible.

Before you go live

Every integration can be validated in a staging environment before production traffic is processed.

Free

Up to 1 week of staging

A full staging setup — using your actual Healthie credentials against a sandboxed target — is provided at no charge for up to 7 consecutive days. This is the recommended path for all new integrations.

Paid

Extended testing

Testing periods beyond 7 days can be arranged. Cost depends on the duration and the volume of data processed during testing. Contact us to get an estimate before committing.

Recommended: Use the free staging period to validate field mappings, edge cases, and AI instruction behaviour on a real subset of your data before scheduling production runs.

What affects cost

Pricing is variable and calculated based on your specific configuration. We do not publish a fixed price list. After an initial scoping call we will provide a tailored quote.

Primary cost drivers

  • Total volume of patient records processed per run
  • Number of connected third-party systems
  • Batch schedule frequency — hourly costs more than weekly
  • Complexity of transformation and business logic
ℹ️

To discuss your specific requirements and receive a tailored quote, get in touch via our main page. We respond to all inquiries within one business day.