Luma breaks an answer into individual claims and checks each one against the primary literature. Every claim is grounded in a real PubMed citation, or honestly flagged. A plain model gives you the same fluent answer, but it cannot tell you which references it made up.
The live engine works across every medical specialty. Ask any biomedical question, or start from one of the examples below.
# Ask a biomedical question. See which claims are real. Live engine, every specialty. Luma breaks an answer into individual claims and checks each one against the primary literature. Every claim is grounded in a real PubMed citation, or honestly flagged. A plain model gives you the same fluent answer, but it cannot tell you which references it made up. The live engine works across every medical specialty. Ask any biomedical question, or start from one of the examples. ## What you can submit - A biomedical question, typed or pasted. - An AI-generated answer you already have in hand. - An attachment: PDF, document, image, or text. Files can be dropped anywhere on the page. ## What Luma does with it 1. Reading the question 2. Retrieving from PubMed 3. Checking each claim against its source 4. Scoring confidence Each claim comes back either grounded, with the PubMed record and a confidence score, or flagged as unsupported. ## Side-by-side comparison Comparison is off by default, so a run makes no third-party call. Turn it on and Luma also fetches an unaided answer from ChatGPT to the same question, then audits that answer against each paper it cites. Citations that are fabricated, or real but unsupportive of the claim, are marked as such. An answer that cites no sources at all cannot be checked against the literature. ## Worked examples ### Cardiology - ACE inhibitor cough: Do ACE inhibitors cause a dry cough, and what else is first-line for heart failure with reduced ejection fraction? - Spironolactone in HFrEF: Is spironolactone recommended in heart failure with reduced ejection fraction, and what monitoring does it require? - Beta-blocker mortality: What is the mortality benefit of beta-blockers in chronic heart failure with reduced ejection fraction? ### Oncology - Trastuzumab in HER2+ breast cancer: Does adjuvant trastuzumab improve survival in HER2-positive early breast cancer, and what is a key cardiac risk? - Checkpoint inhibitors in melanoma: Do checkpoint inhibitors improve survival in advanced melanoma, and what are common immune-related adverse events? - Adjuvant chemo in colon cancer: Does adjuvant oxaliplatin-based chemotherapy improve survival in stage III colon cancer? ### Neurology - tPA in acute stroke: Is intravenous tPA effective for acute ischemic stroke within 4.5 hours, and what is the main risk? - Levodopa in Parkinson's: Is levodopa the most effective symptomatic treatment for Parkinson's disease, and what is a common long-term motor complication? - Disease-modifying therapy in MS: Do disease-modifying therapies reduce the relapse rate in relapsing-remitting multiple sclerosis? ## Before you rely on it Luma is a research and drafting tool, not medical advice. Do not submit protected health information or personal patient data. See the [Terms of Service](/terms) and the [Privacy Policy](/privacy). [Back to Luma](/)