Dermatology AI for virtual primary care and GP networks
A skin check patients can run the moment they open the practice's app or website, with an informational suggestion in under a second.
Used by leading primary care networks and health systems
CE-marked Class I medical device (MDD), MDR Class IIa in progress · MHRA device registered (UK) · FDA Breakthrough Device Designation, clearance pending
What the API returns
One call, one image, one structured response. Everything below is available from the first request.
Five ranked condition suggestions per image, each with a confidence score
70+ conditions across nine dermatology categories: benign, pre-malignant, malignant, inflammatory, infectious, genodermatoses, pigmentation, hair and nail
A response in under a second
No identifiable patient data stored. Images are processed and anonymised
A standard REST call and a JSON response, with no SDK lock-in
ICD-10 reference codes alongside each result, for information only
Full white-labelling. No Autoderm branding is visible to patients unless the provider chooses to show it
Where Autoderm stops
Autoderm analyses the image and returns the ranked list. Everything after that is designed by the provider:
- Whether confidence scores are shown to patients
- Where patients go for further information on a condition
- What the next step offers: a self-care pathway, a booked appointment, a dermatology referral, or nothing at all
- How to provide end user instructions
Autoderm reviews and confirms the compliance of the interface before deployment, so those decisions are made with the provider rather than around it.
The same analysis and interface model runs across Autoderm’s dermatology AI solutions, including dermatology AI for pharmacy and dermatology AI for teledermatology platforms.
What the evidence actually says
Every figure traces to a named study, with full detail on the clinical evidence page.
93%
Top-5 suggestion accuracy
Coachella Study 2025
3M+
API calls, 6 countries
Zero safety incidents since 2018
Post-market surveillance, MHRA/BfArM/FDA MAUDE/ClinicalTrials.gov. Distinct from user-reported dissatisfaction (image quality), not a safety measure.
Bias evaluated across skin tones
Fitzpatrick17K bias evaluation completed across 26 conditions and FST I–VI. No systemic bias detected. Direct same-model cross-population comparisons available (Sweden FST I–II vs. Uganda FST VI on V0.1), with subsequent dataset diversification in V2.x.
Fitzpatrick17K Bias Evaluation, 2024
How the Autoderm API integrates
A dedicated success manager guides the deployment from first call to go-live.
Scope and contract.
Custom pricing and contract terms, set for the network's clinical governance and procurement requirements
Onboard and integrate.
Hands-on integration assistance builds the check into the practice's app or website
Launch and support.
Priority support and custom SLAs carry through go-live and beyond
Regulatory & data
protection standards
Autoderm is developed with strict adherence to European and international requirements for safe, privacy-preserving AI in healthcare.
Common questions about the virtual primary care integration
Is a GP review required?
No. The suggestion can stand alone, or route to booking a GP appointment, a self-care pathway, or a referral. None of these is required by default, and Autoderm doesn’t make that decision.
Does staff need training to use it?
No. The check is self-service for the patient, on the app or website. There is no change to clinical or admin workflows.
How is patient data handled?
No identifiable patient data is stored. Images are processed and anonymised, and the service is GDPR-compliant by design.
How is pricing structured for a network-wide rollout?
Pricing is scoped to the deployment, based on API volume and the contract terms the network needs. There’s no fixed rate card at this tier; the specifics are set with a dedicated success manager as part of onboarding.