About Webspiracy
A sourced encyclopedia of conspiracy theories, government programs, and scandals, mapped by evidence and connection.
How entries are made
Every entry on this site is researched and written by an AI model, given a theory or program to investigate. It searches for real sources, drafts a summary, identifies the people, organizations, and programs involved, and rates each individual claim's evidence status — confirmed, disputed, debunked, or unverified — along with an overall credibility rating for the entry as a whole. Nothing here is hand-written or human-fact-checked beyond the occasional spot correction.
When a person or organization is selected in an entry's connection graph, its card can show a portrait — either pulled automatically from that entity's Wikipedia page or added manually. These are a visual aid, not a verified identification; on rare occasions an automatic search can surface the wrong person entirely, especially for a common name.
AI hallucinations happen here
Despite instructions to only include verifiable claims and cite real sources, the model can still get things wrong — inventing a detail, misattributing a fact, or occasionally stating something that isn't true at all. This isn't hypothetical: an earlier version of one entry on this site stated a CIA report was released in a year that hasn't happened yet. It was caught and corrected, but others may not have been.
Treat every entry as a research starting point, not a final source. Check the linked sources before relying on or repeating any specific claim.
Ratings, explained
Credibility status(shown on each entry) is the model's overall read on the theory as a whole — documented, disputed, debunked, or unverified. Evidence status(shown on each individual claim) is the same scale applied to just that one claim, since a single entry can mix well-documented facts with disputed allegations. Both are conservative by design — the model is instructed to prefer “disputed” or “unverified” over “confirmed” whenever there's genuine uncertainty.
Voting
Priority voteslet signed-in visitors vote on which not-yet-generated theory they'd most like to see written up next — votes just reorder that list, they don't trigger generation automatically.
Provider voteslet you pick which AI provider you'd prefer generate a given theory (Claude, GPT, or Gemini). This site is built to support multiple providers, but generation currently only runs on Claude — provider votes are a preference signal for future generations, not a control that changes anything today.
Support
Every entry here is AI-generated research, and that costs real API calls to produce. If you've found the site useful, you can buy me a coffee — it goes straight toward keeping more entries coming.