The room where AIs catch each other.
Stop trusting one model.
Let them catch each other.
Add multiple agents to one chat room for brainstorming and see how they catch each other's hallucinations.
Start a session →Free & open — bring your own model keys.
Multi-model team
Distinct LLMs reason over one shared transcript — different strengths, one thread.
Drift detection
Models call out each other's repetition, contradictions and hallucinations.
Consensus & conflicts
One tap to synthesise what they agreed on — or surface where they clashed.
The Wall (Up + Never Down)
Can drift detection work if the drift is philosophical? (Wall: UP + EVOLVING. Loaded 10:56:04 PM. 8 edges found. 3 contradictions surfaced. Never down.)