Validate

Trust, evidence, gravity.

Score seriousness, reliability and application-readiness. A composite gravity score the citation crowd never touches, plus a plain decision on whether to rely on the paper at all.

See it live on the origins of AI → Start your own field

[ THE JOBS IT DOES ]

[

When I decide whether to build on a paper, I want a gravity score that ignores the citation crowd.

[

When two papers disagree, I want to know which evidence is actually stronger.

[

When I lean on a citation chain, I want each link's confidence — not a count.

[ WHAT'S REAL · WHAT'S NEXT ]

Live today, and what compounds next

◆ Live now

Composite gravity score (Paper Weight) from 14 intrinsic dimensions.
A decision band per paper: use confidently · use cautiously · needs replication · hypothesis-generating only · do not rely on yet.
The sub-score vector, with the evidence / rigor / gravity / reach rationale.
Claim-level confidence — OnCite verdicts: grounded, citation black hole, circular.
Contradiction detection on the citation edges.

○ On the roadmap

Overclaim detector and replication-proximity index.ROADMAP
Sample-adequacy and bias / confounder scoring as standalone dials.ROADMAP
Evidence-maturity timeline; persona-specific rubrics per paper type.ROADMAP

[ OUTPUTS ]

Gravity leaderboardPer-paper decision bandPer-edge extension score

[ WHO IT'S FOR ]

PIs & lab headsCorporate R&D due-diligenceVCs & funders

This one's live. Watch it read a real field.

The origins-of-AI walk updates hourly — or point it at your own field in minutes.

Open the live demo → Start a field

[ SAME ENGINE · OTHER DOORS ]

One walk of your field feeds every product. Validate connects most closely to:

Paper Twin Understand one paper deeply. Map The shape of the field. Gap Holes become hypotheses. Radar The strategy behind the science.

Part of the Think North platform — like CodeNSM watches your codebase, Re: Sach watches your field. Clarity. Before Commitment.