VoiceClinic Review
A Node.js voice agent that conducts bilingual pre-consultation intake and generates a structured audio podcast for physician review.
Each review covers deployability, value versus commercial alternatives, and privacy model. Tools that can run locally were started and exercised; mobile or backend-dependent tools were assessed from published builds, source code, and deploy guides. Ratings reflect what we were able to verify.
VoiceClinic (Clinical Intake) — AI Pre-Consultation Voice Agent
A Node.js voice agent that conducts bilingual pre-consultation intake and generates a structured audio podcast for physician review. Last updated: 2026-06-21.
One-sentence verdict: A production-oriented voice-intake prototype with Docker, JWT dashboard, and integration tests — a solid scaffold, but not HIPAA-ready out of the box.
What the System Is
VoiceClinic — the repository is published as clinical-intake — is a Node.js / Express application that automates pre-consultation clinical interviews. The patient speaks into a browser microphone; Groq transcribes the speech, an LLM (DeepSeek via an OpenAI-compatible client) selects the next clinical question, and ElevenLabs speaks it back. After the encounter, the system compiles a downloadable MP3 "podcast" summary and exposes a JWT-secured physician dashboard.
The deploy guide reports the service was started on port 5005.
| Key data | |
|---|---|
| Category | Clinical Intake |
| Language | Node.js / Express |
| License | MIT |
| Self-hosted | Yes |
| AI providers | Groq STT, DeepSeek LLM, ElevenLabs TTS |
| Database | File-system sessions; no persistent relational DB |
| Auth | JWT for physician dashboard |
| Tests | 19 integration tests (documented as passing) |
How to Install and Deploy
The repository includes a Dockerfile and docker-compose.yml. The documented flow is:
cd /data2/docker/going-global/repos/clinical-intake
cp .env.example .env
# Add GROQ_API_KEY, LLM_API_KEY, ELEVENLABS_API_KEY
docker compose up -d
Alternatively, the deploy guide uses npm directly:
cd /data2/docker/going-global/repos/clinical-intake
npm install
OPENAI_API_KEY=sk-xxx GROQ_API_KEY=gsk-xxx PORT=5005 npm start
Then open http://localhost:5005 for the patient intake, and http://localhost:5005/login.html for the physician dashboard.
How to Test
The documented test flow is:
- Open the patient intake URL and start a voice encounter in English or French.
- Answer the spoken clinical questions.
- Wait for the encounter to complete and verify that a podcast / MP3 summary is generated.
- Log in to the physician dashboard with the configured credentials and review session history.
- Run
npm testto execute the 19 integration tests, which spawn a local Express server with mock API credentials.
Privacy & Compliance
VoiceClinic is not HIPAA compliant as shipped. The README describes an "anonymous by design" flow with no patient identifiers stored and transient audio processing, but there are no documented BAAs, audit logs, encryption at rest, or access-control policies. Every voice request is sent to Groq, the LLM provider, and ElevenLabs. Do not use it for real patient data without a compliance review.
VoiceClinic vs Commercial Clinical Documentation
| Dimension | VoiceClinic | Commercial Scribes (e.g., Nuance DAX, Ambience Healthcare, Abridge) |
|---|---|---|
| Cost | Free / self-hosted + API usage | Subscription per clinician |
| Deployment | Docker one-command or npm | Cloud SaaS or enterprise install |
| Voice pipeline | Groq + DeepSeek + ElevenLabs | Proprietary ASR/NLU/TTS |
| EMR integration | None; MP3 + dashboard only | EHR-integrated notes |
| Bilingual support | English / French | Varies by vendor |
| Security posture | JWT login, no documented HIPAA controls | BAA, audit logs, encryption |
| Customization | Full source code; question files are plain text | Vendor-controlled |
| Clinical validation | None claimed | Often clinically reviewed |
Who Should Use It
- Developers building a voice-first intake prototype.
- Teams that want a Dockerized starting point with STT/LLM/TTS wired together.
- Organizations that need a bilingual (EN/FR) intake demo and are willing to add compliance controls themselves.
Who Shouldn't Use It
- Clinics needing HIPAA-compliant intake without additional engineering.
- Practices that require EHR integration or structured note export.
- Users expecting a commercially supported medical device.
FAQ
Is VoiceClinic free?
Yes. The code is MIT-licensed. You pay for your own server and any Groq, LLM, and ElevenLabs API usage.
What languages does it support?
The checked-in question files support English and French clinical questionnaires.
Does it store audio recordings?
The README states audio is processed transiently and only the final structured podcast persists. Verify this against the deployed configuration before using with sensitive data.
Verdict
VoiceClinic is the most deployment-ready voice-intake prototype in this batch. It has Docker, a test suite, a physician dashboard, and a clear STT→LLM→TTS pipeline. The main gaps are compliance (no HIPAA controls), no structured data export beyond the podcast, and reliance on third-party API keys. It is a strong scaffold for a development team, not a drop-in clinical product.
Ratings: Deployability 4/5 · Value vs Commercial 3/5 · Privacy Compliance 2/5