Knowledge Graph & Causality Studio (KGG)

Membership required. $3.99/month. Cancel anytime.

Every study rests on a causal model — and that model is almost always trapped in prose, where its flaws stay hidden. KGG pulls it into the light: upload a paper, protocol, dataset, or diagram and render the causal network you are actually claiming — what causes what, where pathways converge, where feedback loops turn, and where the structure will betray you. It flags the colliders that manufacture false associations, the common causes that masquerade as effects, and the bottlenecks every path must cross. Spin it in 3D, send a causal pulse downstream, right-click any node to read the inference it implies and the trap it hides. Causal honesty made visible — in a tool you can use in an afternoon.

Upload anything → get a knowledge graph / causality network

• Research papers & preprints (PDF), abstracts, systematic reviews & meta-analyses
• Study protocols and your own manuscript drafts
• Case reports & de-identified case series (built-in, in-browser de-identification)
• Datasets & lab results (CSV / XLSX)
• Grant proposals & NOFOs — the theory of change made explicit
• Any document (DOCX, TXT, Markdown), web pages, news articles & URLs
• Books & chapters, lecture notes, transcripts, interview / qualitative data
• Policy papers, white papers, contracts; business plans & strategy docs
• Incident reports & post-mortems (root-cause / blast-radius)
• Existing diagrams & model images (PNG / JPG) — read into an editable, analyzable graph
• Slide decks, hand-drawn causal sketches, or just pasted text

What it solves

• Causal claims are buried in prose — KGG makes them visible and clickable.
• Inference traps stay invisible until peer review — KGG flags colliders, confounding (Table-2 fallacy), and bottlenecks before they cost you.
• Study design gets the adjustment set wrong — seeing the graph tells you what to condition on and what never to touch.
• Literature synthesis is slow — upload several papers, get one integrated causal picture.
• Observational claims get overstated — KGG shows when the structure cannot support the claim.

Roles and hazards shown by KGG are implied by graph topology — position and role, not proof of the arrows or their magnitude.