We make money when you save time—not when we harvest your data.
Most AI companies have a fundamental conflict: the more data they collect, the more valuable their AI becomes. That creates pressure to retain your data, aggregate it, and train models on it.
NovaScribe™ has no such conflict. We’re a medical transcription service, period. We make money when you save time—not when we harvest your data.
That means:
“This isn’t a policy we adopted for marketing. It’s the only way we’d want our own patient data handled.”
| Your Data | What We Do | Why |
|---|---|---|
| Audio recordings | Deleted within 24 hours | We don't need it after transcription |
| Transcripts | Stored encrypted in your account | Accessible only to you |
| AI training | Never used | Not our business model |
| Third-party sharing | Never | No exceptions |
Full administrative, physical, and technical safeguards
Business Associate Agreement for all accounts
Data encrypted at rest and in transit (TLS 1.3)
AWS US-East-2, SOC 2 certified data centers
NovaScribe has completed third-party penetration testing to validate our security controls.
Conducted by Beagle Security
Zero critical, high, or medium vulnerabilities
Security controls validated
Payment security validated
NovaScribe’s evidence search intelligently searches clinical practice guidelines, systematic reviews, and randomized controlled trials to answer your clinical questions with AI-synthesized summaries.
Unlike general-purpose AI tools that may generate plausible-sounding but fabricated citations, every reference returned by NovaScribe’s literature search is checked against PubMed and DOI registries for existence before it’s shown to you. Review each citation before relying on it.
Every cited reference is checked against PubMed and DOI registries for existence, so you can independently verify each source.
“AI-generated summaries are drawn from published literature and may contain errors. We always encourage physicians to verify citations before clinical use -- and we make that easy: NovaScribe checks each citation against PubMed and DOI registries for existence, and you review it before relying on it.”
We deliberately chose HIPAA-compliant enterprise platforms over direct AI vendor APIs:
| AI Processing | Platform | BAA With |
|---|---|---|
| Note generation | AWS Bedrock | Amazon |
| Clinical analysis | Microsoft Azure | Microsoft |
| Speech recognition | Google Cloud |
Why this matters: Direct API access to AI vendors often lacks HIPAA compliance. By routing through enterprise platforms like Microsoft Azure and AWS Bedrock, we get Business Associate Agreements with trusted enterprise vendors, zero data retention guarantees, and contractual prohibition on training.
Your data passes through, gets processed, and is gone. It never becomes part of anyone’s training dataset—not ours, not Microsoft’s, not Amazon’s.
We're happy to answer any questions about our security and privacy practices.
10 notes per month free. No credit card required.