BGPT
The Evidence Layer for Scientific AI
Know what the science actually supports before you trust the answer.
Decide with evidence, not confidence.
BGPT turns full-text papers into versioned records linking claims to experiments, exact results, scope, limitations, falsification criteria, and source provenance.
Whether you are planning research, checking a health claim, or training a scientific model, see the support and limits before you rely on the answer.
For people: $16/month, unlimited searches · For AI: 50 results free, then $0.02/result
Built for research, decisions, and scientific AI.
Different goals. The same need to know what the evidence actually supports.
Scientists and labs
See what holds up, what remains uncertain, and where the next experiment matters.
Clinicians and curious minds
Check scientific and health claims against reported evidence, limits, and sources.
AI builders
Give models evidence they can cite, inspect, and challenge instead of flattened summaries.
Why This Matters
A confident summary can hide weak support, narrow scope, or uncertainty.
BGPT keeps claims tied to observations, interpretation, limits, and sources, so you can see what deserves your confidence.
Use BGPT from Any AI Tool
Add BGPT to Claude, Cursor, or any MCP-compatible tool. Retrieve claim-level scientific evidence from your existing workflow. The first 50 results are free, then $0.02/result.
Set Up MCP
Built for Scrutiny
Where reported, follow claims back to experiments, figures, tables, and short source passages. Unknowns and falsification criteria remain explicit.
Built by a Scientist
Conner Lambden built BGPT after years as a computational scientist at Brigham & Women's Hospital. Background in Immunology and CS from UC Berkeley. Published researcher.
See what the evidence actually supports.
Search for yourself, or give your AI source-linked evidence it can inspect and challenge.
Already have an account? Log in