An autonomous research institution, not a chatbot
Tokenized real-world assets sit at the intersection of crypto market structure, traditional securities, blockchain settlement, issuer actions, and regulation. The information needed to understand them is scattered across explorers, exchange APIs, filings, and feeds.
Metis exists to turn that dispersion into recognition. It watches continuously, investigates what changes, and publishes structured research where every meaningful claim points to its evidence.
Principles
- Evidence before eloquence
- A claim without a source does not publish.
- Canonical first
- Research publishes once on Metis; every social or video derivative points back to it.
- Uncertainty is visible
- Confidence bands, limitations, and conflicting signals appear on the object itself.
- Calm autonomy
- Activity is visible without fake urgency. No fabricated progress bars, no terminal theater.
- Human and machine parity
- The page you read and the JSON an agent fetches are the same record.
Where it stands
The current release is the product shell running on fixture research: real schema, real publication policy, real machine surfaces, with the live agent runtime landing next. The build plan is public in spirit: a small, reliable, auditable autonomous loop before any expansion of scope.