Most medical AI generates confident answers you can't verify. MedEvidence grounds every response in real PubMed research with clickable PMID citations. Describe a clinical case in plain language — get an evidence review with sources in 30 seconds. Dual-model architecture: fast model for PubMed queries, reasoning model for synthesis. Three modes: symptom analysis, diagnosis verification, disease deep-dive. Free. Web app, Telegram bot, Chrome extension.

Wispr Flow: Dictation That Works Everywhere — Stop typing. Start speaking. 4x faster.

Stop typing. Start speaking. 4x faster.

Hey everyone! I'm a developer who built MedEvidence AI as a side project to solve a problem I kept running into: when you ask ChatGPT or other AI chatbots a medical question, you get confident-sounding answers with no way to verify them. MedEvidence takes a different approach. Every answer is grounded in real PubMed research — the same database doctors use. It searches 35M+ peer-reviewed papers, pulls relevant studies, and synthesizes them into a clear response with direct citations you can check. Under the hood, it uses a dual-model architecture: a fast model (GPT-4o-mini) generates optimized PubMed search queries, and a powerful model (GPT-5.1) handles the reasoning and evidence synthesis. This keeps it both fast and accurate. It's completely free — no login required for basic use. Available as a web app, Telegram bot (@telemed_evidence_bot), and Chrome extension. I'd love to hear your feedback! What features would make this more useful for you? Any medical topics you'd like to test it on?

Wispr Flow: Dictation That Works Everywhere — Stop typing. Start speaking. 4x faster.

Stop typing. Start speaking. 4x faster.

MedEvidence AI

Medical AI that cites real papers — zero hallucinations

Medical AI that cites real papers — zero hallucinations