On 27 April 2026, OpenEvidence became unavailable in the European Union and the United Kingdom. The company cited mounting regulatory uncertainty, including the EU Artificial Intelligence Act. A subsequent commentary in The Lancet Regional Health - Europe described the event as a warning about the need for clearer medical-AI governance.
The immediate problem for clinicians is practical rather than political. OpenEvidence had made a fast, cited response to a clinical question feel normal. Published reporting described roughly 18 million clinical consultations each month before the withdrawal. The questions that created that habit still arise on rounds, in clinics, and during research.
What European clinicians lost
The useful part of the experience was not simply an answer generated by a language model. It was the combination of retrieval, readable synthesis, and references that a clinician could inspect. When that combination disappears, users may fall back to general-purpose chatbots that are not designed to show where a medical claim came from.
Why OpenEvidence stopped working, in plain terms
If you found this page because OpenEvidence suddenly stopped loading or your login no longer works, here are the short answers.
- Why is OpenEvidence not working in Europe?
- OpenEvidence withdrew access for users in the European Union and the United Kingdom on 27 April 2026, citing regulatory uncertainty that included the EU Artificial Intelligence Act. The withdrawal applies to existing accounts as well as new registrations, so this is a deliberate geoblock rather than a technical outage.
- Can I use OpenEvidence with a VPN from Europe?
- A VPN might load the website, but it does not solve the real barrier: clinician verification requires a US National Provider Identifier (NPI), which European clinicians do not have. Circumventing a geoblock is also likely to breach the service's terms of use. In practice, European clinicians need a tool that verifies European professional credentials.
- Will OpenEvidence come back to Europe?
- There is no announced timeline for a return. Commentary in The Lancet Regional Health - Europe framed the withdrawal as a governance problem, which suggests access may not resume until European medical-AI rules feel more predictable to US vendors.
- What should European clinicians use instead?
- Several tools serve European clinicians, including CliniAtlas, which is EU-hosted and free for verified professionals and students. Whichever you evaluate, the checklist below matters more than any brand name: inspectable citations, European sources, EU data processing, and a published evaluation method.
Six questions to ask any replacement
Several products now describe themselves as alternatives for European users, CliniAtlas included. The following checklist is more useful than a feature comparison:
- Can you open and inspect the source behind a material claim?
- Does the interface distinguish guidelines, regulatory records, reviews, trials, and preliminary evidence?
- Does retrieval include European regulatory sources and guidelines where they are relevant?
- Where are application workloads and stored user data hosted?
- Is the intended-use boundary stated clearly, especially for patient-specific questions?
- Can the system abstain or expose limitations when retrieval is weak?
Where CliniAtlas fits
CliniAtlas is built in Split, Croatia, in collaboration with the University of Split School of Medicine. The application, language-model inference, and stored service data run in AWS eu-central-1 in Frankfurt. On-demand searches call public evidence services including PubMed and Europe PMC, and the product also uses curated European guideline and EMA medicine registries.
Answers are designed around inspectable citations and a reference panel that labels source type and evidence stratum. The product works in English and Croatian today. It is intended for literature search, education, and professional evidence review, not diagnosis or patient-specific treatment selection.
CliniAtlas is younger than the platform Europe lost, and we do not present an early product as settled clinical infrastructure. Our aim is narrower: build a transparent evidence workspace for European professionals, measure it honestly, and improve it with clinical and academic partners.