From raw signal to reviewed brief

Sources come in

Market data, on-chain/DeFi data, and news are collected from multiple independent providers.

Related stories are grouped

Coverage of the same underlying event is clustered together, and near-duplicate coverage is flagged rather than treated as separate confirmation.

Claims are extracted

Each source is read for its specific factual claims — what happened, to which asset, and when — rather than treated as a single opaque headline.

Sources are weighed for independence

A claim repeated by five outlets syndicating the same wire story doesn't count as five confirmations — source independence is checked before anything is treated as well-corroborated.

The event is classified

Each story is placed into one of a fixed set of event categories (for example, a regulatory development or a protocol security event) — never left as free-form guesswork.

Market reaction is checked

Price and volume behavior around the event is examined to see whether the market is actually confirming the story. When that confirmation isn't available, the brief says so explicitly rather than assuming it either way.

Relevance to your portfolio is scored

The event is weighed against your actual holdings and exposure — a story about an asset you don't hold never competes for attention with one that does.

A label is assigned

Based on the evidence gathered, each affected asset receives one of six fixed labels, along with a confidence score. Missing or thin evidence pulls the result toward a lower-confidence, more cautious label — it is never treated as a green light.

A draft brief is assembled

The labels, evidence summaries, and narrative are assembled into a draft brief, checked against a set of citation and compliance rules before it goes any further.

A person reviews it before you see it

Every draft brief is reviewed by a person before it is approved and made visible to you or sent by email — the system alone does not get the final say on what reaches your inbox.