Mirror lets institutions jointly catch trade-based money laundering across their walls — without pooling a single customer record — and turns every decision, including the clears, into a signed record a regulator can inspect.
Not zero-knowledge. Not MPC. Blind to the operator by construction — and we say exactly where that ends.
Trade-based money laundering is one of the largest and least-tooled laundering channels in the world, and the mechanism is simple: declare a 1,000-unit shipment at $300 a unit when it is worth $100, and $200,000 has moved out of the country under cover of a legitimate-looking trade. It stays invisible for one structural reason.
Members agree a public reference price-band table per product. Mirror invents no prices — it only ever uses the table the parties supply.
Inside each institution a raw value becomes a band, and only a salted commitment of that band leaves the building. The raw number never crosses the wall.
The operator compares commitments only. It is structurally unable to read a value, because a value is never in its input.
The verdict is signed and canonical. A flag routes to review; a clear stands as the defensible reason it was cleared.
Single-institution monitors cannot see across the wall by architecture. The whole value here is the join between institutions — and that is a network an incumbent’s customers cannot rebuild alone.
FinCEN 314(b), the EU’s AMLA framework and privacy-enhancing-technology sandboxes at the FCA and ICO exist precisely to enable sharing of this shape. Mirror is the tool those rules anticipated.
Most AML tooling is priced on the alerts it raises. Mirror also gives you the means to defend the ones you didn’t raise — one of the hardest questions in any examination, answerable offline from the signed record.
A compliance team should never have to reverse-engineer what a vendor actually guarantees. These are the limits, written the same way in the product, the documentation and the pilot agreement.
Salted band-commitments blind the operator, that lacks the per-shipment secret. They do not hide the band from the counterparty who shares that secret. The threat model is “don’t pool raw trade data centrally” — not “hide the two legs from each other.”
Mirror never asserts laundering; it surfaces structure. A clear is a defensible record of why a trade was not escalated — never a guarantee of innocence.
A missing leg, an unbandable value or an unreadable commitment is undetermined — and can never clear. Ambiguity escalates, by construction.
Land with a correspondent-banking pair — or a bank and a customs or trade-data source — who each hold one leg of the same real shipments. Run a month of historical trades through Mirror. One signed divergent verdict on a mispricing both sides individually missed is the whole proof.
Pilot enquiries · [YOUR PILOT CONTACT EMAIL] · [YOUR COMPANY ENTITY]
Working name — after central-bank mirror statistics.