Facts, fiction, and bounties.
Every submission gets triangulated by two independent LLMs (Claude Sonnet + Gemini) plus Google grounding. Consensus → FACT or FICTION and AEGIS mints to the submitter automatically. LLM disagreement → BOUNTY — anyone can resolve it. Every 256 verified truths get sealed into a permanent Merkle-signed block (further down this page). Spot a mistake? Dispute — the LLMs re-run against your reasoning and the chain corrects itself.
What do these labels mean?▾
Every settled row becomes a labelled example in an open training dataset that teaches future AI models historical fact-patterns.
Every 256 truths above get bundled, Merkle-rooted, HMAC-signed, and linked to the previous block via previous_hash. Anyone — including a competing AI — can query /api/truth-blocks/{id}/proof/{citation_id} to prove inclusion in O(log n) hashes. Chain integrity: verified ✓.