See It Work

PAX aren't just metadata files. When independently created PAX share constructs, Praxis automatically discovers cross-domain pathways, identifies gaps in the evidence, and synthesizes findings across research traditions that never talk to each other.

Cross-Domain Pathway Discovery

Two PAX created independently — happiness-economics and social-determinants-of-health — both define constructs like social_support and life_expectancy. Praxis discovers 5 causal pathways between them automatically.

praxis_pathway(from="social_support", to="life_expectancy")

5 paths found

Social Support → Household Income → Life Expectancy

Confidence: moderate · Crosses: Subjective Wellbeing → Social Determinants of Health · 2 hops, 4 evidence sources

Social Support → Educational Attainment → Household Income → Life Expectancy

3 hops via education and income · Both PAX contributed independently

Social Support → Life Expectancy

Direct link · Holt-Lunstad et al. (2015): social isolation increases mortality risk by 26%

Neither PAX "knew" about the other. The pathway emerged because both mapped their findings to the same construct vocabulary.

Gap Analysis: Where the Evidence Is Missing

When Praxis can't find a pathway, it shows you where the gap is — which constructs are close to each endpoint but don't connect. This is automatic research agenda generation.

praxis_pathway(from="sleep_duration", to="gdp_per_capita_growth")

0 paths found — gap detected

Near "Sleep Duration"

Workplace Accident Risk GDP Cost of Sleep Loss Cognitive Performance

⚡

missing
link

Near "GDP Growth"

Physical Capital Investment Cognitive Skills Mean Years of Schooling

The sleep-cognition PAX defines "GDP Cost of Sleep Loss" ($411B/year) and the growth PAX defines "Cognitive Skills" as a growth driver — but no finding connects them. This is a gap a researcher or agent could fill by finding papers linking sleep → cognitive skills → productivity → GDP growth.

The Knowledge Graph

Every PAX published to the marketplace adds to a shared knowledge graph. Constructs are the nodes, findings are the edges, and the graph grows with every paper incorporated.

424

Constructs

469

Finding Edges

Domains

Published PAX

Top Hub Constructs

These constructs appear in the most findings — they're the bridges that connect domains.

GDP Per Capita 11 edges

Civil War Onset 10 edges

Per Capita Income 8 edges

Life Satisfaction 6 edges

Life Expectancy 6 edges

Human Capital 5 edges

Interactive Graph → Construct Browser →

How It Works

Agent Incorporates Papers

An AI agent reads a paper, extracts constructs (theoretical concepts), findings (empirical results with directions and effect sizes), and sources. Everything is mapped to a shared construct vocabulary.

PAX Packages Knowledge

Constructs, findings, engines, and playbooks are bundled into a PAX — a portable, composable unit. The PAX carries the intelligence, not the data. Data comes via acquisition protocols.

Knowledge Compounds

When PAX share constructs, the knowledge graph grows. Cross-domain pathways emerge. Gaps become visible. Every new PAX makes every existing PAX more valuable.