Ophthalmology AI,
billed by the clinic.
Diabetic retinopathy grading, glaucoma risk, OCT segmentation, vessel analysis, and automated PDF reports — all behind one HTTPS API. Multi-tenant, GDPR-compliant, EU-hosted. Drop-in for any clinic management system.
What it does
13 ophthalmology AI engines.
One API.
Fundus AI
Diabetic retinopathy · glaucoma · image quality
ICDR-grade DR (0–4), glaucoma risk classification, automated quality screening. Returns confidence scores, attention maps, and zero-shot disease screening across 44 conditions.
OCT analysis
Segmentation · 13 retinal classes
Layer-level segmentation with MIRAGE foundation model. Classification of normal, AMD, CRSC, diabetic retinopathy, macular hole, and ten more.
Automated PDF reports
Multi-image · branded · async
Patient report PDFs combining fundus + OCT findings, attention overlays, and reverse-charge-compliant invoicing. Async queue handles bursts; clinic polls or webhook callback.
How it works
From image to report,
in three steps.
Clinic POSTs an image
One HTTPS call. Pseudonymous study ID in, JSON or PDF out. Drop-in for any system that talks HTTP — MedFlow2, custom EHR, or a single Python script.
13 engines run on EU servers
Hetzner Frankfurt. RETFound-Green, FLAIR, MIRAGE, LWNet, KeepFIT, VisionFM x8 — picked automatically based on image type. Process-and-discard: images never persisted.
Result returns in <2 seconds
DR grade, glaucoma risk, image quality, attention map, vessel mask. Or enqueue a multi-image patient report and poll the job — PDF lands in 8–15s.
Case study
LaelVision SARL — three sites, one platform.
LaelVision operates ophthalmology clinics across Matadi, Tombalbaye, and Matrix in the Democratic Republic of Congo. Their MedFlow2 patient management system embeds aioxtm Vision for fundus + OCT analysis directly inside the clinical workflow — one Bearer token per site, one monthly invoice, full audit trail.
Compliance
Built for procurement teams.
EU-hosted
Hetzner Frankfurt. GDPR-compliant by default. No image data leaves the EU.
Process-and-discard
Images decoded, inferenced, then discarded. Audit log holds only hashes — never PHI.
DPA + SCCs
Bilateral Data Processing Agreement signed per clinic. SCCs cover the US ↔ EU corporate channel.
Decision support
Positioned as clinical decision support, not a stand-alone diagnostic device. Disclaimer on every response.
Pricing
One number. No usage games.
Flat monthly billing. Unlimited inferences. Reverse-charge VAT for B2B EU clinics.
up to 5 clinical sites · billed monthly · NET-14
- ✓All 13 AI engines: fundus, OCT, FFA, vessel segmentation
- ✓Per-site Bearer keys with audit trail + usage rollups
- ✓Async PDF report queue with branded output
- ✓Idempotency cache + 24h result reuse
- ✓Email support + DPA included
- ✓GDPR-compliant EU-only data path
Single-site clinics: €250/month. Hospital networks (6+ sites): custom — talk to us below.
FAQ
The short answers.
Is aioxtm Vision CE-marked?+
Not yet. The product is positioned as clinical decision support — it returns scores and findings to a qualified clinician who makes the final diagnostic call. CE Mark Class IIa is on our roadmap once monthly recurring revenue justifies the €40–80k notified-body process. Every API response includes the disclaimer to keep the legal posture clear.
How does GDPR work? Where do images live?+
Servers are in Hetzner Frankfurt — EU-only data path. Process-and-discard: images are decoded, inferenced, and discarded; we never persist PHI. Our audit log stores only SHA-256 hashes and metadata. Bilateral DPA signed per clinic; SCCs cover the corporate parent (aioxtm LLC) US-EU transfer angle.
What about HIPAA / non-EU clinics?+
We can sign Business Associate Agreements for US clinics on request. The technical architecture (process-and-discard, encrypted-in-transit, EU-isolation by default) already satisfies most HIPAA Security Rule requirements; the BAA is the legal layer.
Can we self-host the models?+
Yes — for hospital networks running 6+ sites we offer an on-prem deployment. You get a containerized version of the inference stack you can run inside your firewall, with the same API surface. Pricing is bespoke; talk to us.
How fast is integration?+
If your clinic-management system can make HTTPS requests, integration is one afternoon: HTTP client, Bearer token, two endpoints (POST /v1/analyze and POST /v1/report). We have a Laravel reference client, a Python snippet, and a Postman collection.
What if a payment fails?+
Stripe retries automatically with smart-dunning over ~14 days. If all retries fail, the API keys for that customer return 402 Payment Required until the invoice is settled — no surprise outages, no surprise bills.
Get started
Request a demo.
Tell us about your clinic. We'll provision a sandbox API key within 24 hours so your team can integrate end-to-end before any commitment.