Do you use Whisper or commercial ASR?

It depends on your compliance requirements. For HIPAA workloads we typically self-host Whisper so audio never leaves your controlled environment, and we use commercial ASR like Deepgram or AWS Transcribe Medical where a Business Associate Agreement is available. We’ll tell you honestly which trade-off fits your data and budget.

Can you integrate with Epic / Cerner?

Yes. We ship EHR write-back over FHIR R4 and HL7 v2.x to Epic, Cerner / Oracle Health, Athenahealth, and eClinicalWorks, so finalized notes post straight back into the chart. If your EHR exposes an integration path, we can wire it in. If it doesn’t, we’ll tell you the honest workaround.

How do you handle PHI in LLM processing?

We route PHI only through HIPAA-eligible LLM providers: Azure OpenAI and AWS Bedrock, and only under a Business Associate Agreement. No PHI touches a model or vendor that isn’t covered by a BAA, and that boundary is architected into the pipeline, not left to configuration.

What about specialty-specific note templates?

Yes. Primary care, cardiology, dermatology, and other specialties each have different SOAP structures, and we build the templating layer so each note follows the right format. New specialties are added as templates rather than rebuilds.

Do you handle ICD-10 / CPT coding?

Yes. We build coding assistants that propose ICD-10 diagnosis codes and CPT procedure codes as billing-ready output. Coding is always assistive: the clinician reviews and finalizes before anything is submitted, with the audit trail to back it up.

AI Clinical Documentation / Ambient Scribe Development

We build ambient AI scribes, SOAP note generation engines, and ICD-10/CPT coding assistants. HIPAA-architected. EHR write-back ready.

Discuss my project See what we build

What we build

Ambient scribe systems

Real-time audio capture, transcription, and summarization during clinical encounters.

SOAP note generation

Structured clinical notes from transcripts, with specialty-specific templates.

ICD-10 + CPT coding assistants

Diagnosis coding, procedure coding, and billing-ready output, with clinician final review.

EHR write-back integration

Epic, Cerner, and Athenahealth integration for posting finalized notes back to the chart.

Voice-to-text pipelines

Whisper (HIPAA-eligible deployments), Deepgram, and AWS Transcribe Medical.

Clinician review workflow

Note draft → clinician edit → finalize → EHR post, with audit trails on every step.

Doctor using a tablet in a clinical environment

Clinical tools that fit into real workflows

Reference architecture

Audio Input
ASR (Speech-to-text)
LLM Summarization
Structured Output
Clinician Review
EHR Write-back

A typical ambient scribe system: audio from the encounter is transcribed by a HIPAA-eligible ASR layer, the LLM summarizes it into a structured SOAP note, the clinician reviews and finalizes, and the approved note writes back to the EHR over FHIR / HL7.

Integrations shipped across 13+ partners.

EHRs

Epic
Cerner / Oracle Health
Athenahealth
eClinicalWorks

Standards

FHIR R4
HL7 v2.x

Speech-to-text

Whisper (self-hosted for HIPAA)
Deepgram
AWS Transcribe Medical

LLM (HIPAA-eligible)

Azure OpenAI
AWS Bedrock

Audio infrastructure

Twilio Voice (HIPAA)
Real-time streaming

Which engagement fits

Project Build

For a new ambient scribe platform: capture, transcription, SOAP generation, coding, and EHR write-back, built in a focused scope.

Explore Project Build

Embedded Squad

For scaling specialty support: a dedicated team that ships alongside yours as note templates, coding, and integrations expand.

Explore Embedded Squad

Tech Audit

For assessing an existing build: a 5-day diagnostic of architecture, ASR/LLM pipeline, and especially HIPAA and BAA compliance.

Explore Tech Audit

Compliance & BAA posture

Compliance and BAA coverage for AI clinical documentation builds. Status reflects how we engineer the controls into your platform. Formal FDA and state-board determinations stay with your consultants and advisors.
Standard	Status	What we ship
HIPAA	Compliant	Compliant builds with PHI handling, access controls, encryption, and audit logging architected from day one. See /compliance/hipaa.
BAA coverage	Compliant	We architect so every LLM and ASR provider in the path is HIPAA-eligible and under a Business Associate Agreement: Azure OpenAI, AWS Bedrock, and self-hosted Whisper for workloads without a BAA option.
FDA SaMD	In progress	Ambient scribes are generally advisory and not Class II SaMD, but classification depends on intended use. We architect for compliance; a formal determination requires your FDA regulatory consultant.
State medical board rules	In progress	State rules on AI-assisted documentation vary. We build in clinician-review gates and disclaimer language; the applicable board requirements stay with your regulatory advisor.

Free

HIPAA Stance Builder

A guided walkthrough that maps your scribe platform against HIPAA: PHI flows, BAA coverage across ASR and LLM providers, encryption, and audit logging, and hands you a defensible compliance stance.

Build your HIPAA stance

Medical imaging displays in a clinical setting

Engineered for clinical delivery

Patient data, provider workflow, regulatory guardrails

Clinical systems are measured by uptime and auditability, not marketing slides. We build for the reality of busy practices, compliance deadlines, and the patients who depend on both.

Frequently asked questions

It depends on your compliance requirements. For HIPAA workloads we typically self-host Whisper so audio never leaves your controlled environment, and we use commercial ASR like Deepgram or AWS Transcribe Medical where a Business Associate Agreement is available. We’ll tell you honestly which trade-off fits your data and budget.

Building an AI clinical scribe? Schedule a meeting.

We’ve architected ambient scribes, SOAP generation, and coding assistants with HIPAA and BAA coverage built in from day one. Tell us what you’re building.

Discuss my project Back to AI Healthtech