Ask Nuclide

Ask in plain words. Nuclide works out what you are asking and which records you named, then builds the answer from those records, plain English first, each sentence numbered and linked so you can check it at its source. Nothing is generated and nothing leaves your browser.

Looking for unmet needs or candidates rather than a fact about a record? Search the whole universe: every registered product, trial, paper, congress abstract, patent and published unmet need, with ideas computed from the cross-links.

Answers are assembled from Nuclide records by rules, not generated. Each sentence is numbered and linked to the record it came from. Not medical advice.

Try a question

(patient)
(patient)
(clinician)
(clinician)
(analyst)
(analyst)

How it works. Your question is read for what it asks (what something means, how a cancer is treated, whether a product is approved, how it works, side effects, trials, results, a comparison, outlook, who or where, cost) and for the records it names, matched against every Nuclide name and alias. Those records are fetched and a template for that kind of question fills the answer from their fields, plain English first, one citation per sentence. Approval answers start with your chosen region.

What it cannot do. It does not reason, weigh evidence or fill gaps. If Nuclide has no record on the point, it says so and falls back to the closest sentences from word and concept search. The open benchmark scores this method in public.