Reasoning core stays protected
Clients receive only what they need to request and display results; provider secrets and protected instructions stay server-side.
PRIVACY · SAFETY · AGENCY
The tool should increase agency, not become another surveillance layer.
Mīzān keeps sensitive reasoning logic on the protected service, limits what the browser and phone receive, and starts analysis only after an explicit action from you. Safety and privacy are part of the product itself.
No automatic publishing
Mīzān may draft or insert text, but the final platform action remains yours.
No truth authority
Uncertainty and verification limits remain visible instead of being collapsed into a binary verdict.
Trust claims must match reality
Mīzān does not claim deletion, privacy or billing behavior that the backend does not actually provide.