Ask any health question. See what the research actually says.
Peer is now closed. This page is an archive of what we built.
Full methodology shown on every answer
When you ask Peer a question, it goes to the same databases used in clinical research.
Live search across 9 databases and 64 trusted publishers
What happens behind every answer
Every answer passes through the same sequence: search, rank, grade, then write. The evidence grading is deterministic. The AI writes the answer but doesn't control the process.
Parallel database search
9 medical databases run in parallel, plus 64 trusted health publishers via web search in the same step. If one is slow or down, the others keep going.
Semantic re-ranking
A relevance model re-sorts results by how directly they answer your specific question.
Evidence grading
Each study is graded by design quality using deterministic rules the AI cannot influence. Randomized trials rank higher than case reports. The grading happens before the AI ever reads the evidence.
Quality checks
If results aren't relevant enough, the system broadens its search and tries again before writing the answer.
Claim verification
After the answer is delivered, a separate AI re-reads each cited study and checks whether every claim is actually supported. Results feed our quality monitoring.
Independent quality scoring
Every response is scored across multiple dimensions by an AI that didn't write the answer. One factual error means an automatic zero. These scores drive continuous improvement.
These demos show a condensed preview. Real answers are longer, with more citations, evidence grading, and a full methodology breakdown.
Field notes
What people are saying
What I like is being able to see the actual research and citations behind every answer.
I used to spend hours trying to figure out if something was legit. Now I just ask Peer.
I've been using this over OpenEvidence for my research.