Doctorina

Doctorina is an AI health assistant that keeps asking, one question at a time: you feed it symptoms, a voice note, a photo of a rash or a lab report, and it follows up point by point the way an outpatient doctor would, then closes on one of three tiers — self-care, see a doctor, or go to the emergency room now. Its target is not insured users in Europe or North America, but regions where booking a visit is hard, languages are mixed and connections are poor. The official framing is unambiguous: an informational health support tool that does not constitute diagnosis or treatment advice.

Core capabilities

• Question-by-question intake: instead of dumping an answer up front, it adds questions based on your replies and narrows a vague description into a case that can actually be assessed
• Triage that always closes: every conversation ends with an explicit next step — watch it at home, see a doctor, or go to the emergency room now
• Red flags mean care now: when your answers hit a warning sign, it gives direct instructions to seek care immediately or call emergency services
• Multimodal input: text, voice, photos and medical record files all go straight into the conversation
• 90+ languages: a built-in translation step keeps even less common languages on the same intake flow
• Printable clinical summary: export a PDF in one click and hand it to an in-person doctor for a second opinion

Where it fits Symptoms you cannot describe clearly, no idea which specialty to book, a lab report you cannot read, a prescription you want to double-check, or a country whose language you do not speak — sorting information before a visit is its home turf. Emergencies, confirmed diagnoses and medication decisions are out of scope.

What sets it apart Compared with general chat models, it forces every conversation to close on where to go next and ships a built-in red-flag checklist; compared with Ada and K Health, it lacks CE/FDA clearance and a loop with real doctors, but its multimodal entry points and 90+ language coverage are broader. Credibility is still self-reported: the site claims 95% accuracy, while the team's own arXiv preprint (150 synthetic Polish cases) reached only 82% on Top-1, and the 200K consultations, 60 countries and 90+ languages have no third-party verification; the site says data is stored only on European servers and never used for ads, yet the privacy policy includes a US region and lists ad-side processors such as Google Ads and Facebook Analytics.

Editorial take Use it as a navigation aid to care rather than as an AI doctor. For people who struggle to get an appointment it has real value, and the multilingual coverage and exportable summary are especially rare. But the free tier allows only 1 consultation per day and full recommendations require a subscription that renews automatically, so spend the free allowance on two or three conversations before deciding.