Anti-Conspiracy Theorem
A working public lab for inspecting claims by evidence shape, source diversity, falsifiability, and rhetorical pressure.
Analyze
ReadyDrop a document, paste text, or analyze a URL.
Load a document to build the graph.
Claims, evidence traits, and penalties will appear as nodes after the lab scores real text.
CosmosGL claim graph
No document loaded
Average ACC
No run
Strong
0
Mixed
0
Suspect
0
Install the reader, then inspect the trace.
The public lab runs a deterministic browser scorer. The extension uses the same ACC shape and can prepare the local WebLLM model through the owned `/act` route.
NPM install command
npx act-theorem install --out ./anti-conspirarcy-theorem-extensionLocal first
Dropped documents are scored in the browser.
URL mode
The server reads the URL, then returns extracted text.
Traceable
Node details show rules, penalties, and actions.
The theorem is a graph problem.
Claims gain strength when evidence roots are independent, specific, temporally spread, and tied to sources. They lose strength when support collapses into circular citation, unfalsifiable language, or high rhetorical pressure.