About

We are building infrastructure for programmable markets — software that lets parties discover, transact, and settle on assets that have historically been bought and sold through brokers, spreadsheets, and email threads. Our work is rooted in three disciplines we treat as complementary rather than competing.

Trading

We borrow from financial market design: continuous orderbooks, periodic auctions, transparent price discovery, and the operational machinery — state machines, idempotency, audit trails — that lets people transact at speed with confidence. Many markets outside finance, particularly those involving real-world rights, time-bound capacity, or scarce attention, are underserved by today's tools and would benefit from the same primitives.

Web3

We use cryptographic constructs where they reduce trust requirements without adding friction: signed messages, deterministic settlement state, programmable escrows. We don't believe every market needs to run on a public chain, but we believe every market benefits from the operational properties chains gave us — provenance, immutability, verifiable history. Our architecture treats settlement as a first-class concern, not an afterthought.

AI

Modern markets generate more data than humans can act on in real time. We use machine learning for the boring-but-essential work: matching counterparties, surfacing relevant inventory, pricing under uncertainty, detecting anomalies, and reducing cognitive load on participants. We are deliberately conservative about handing decisions to models; AI lives downstream of well-defined market rules, not in place of them.

What we are doing

We are a small team, building patiently, with a bias for systems that work for years rather than products that look impressive for weeks. We are not announcing specifics yet — the right time to talk about a market is when participants are using it, not before.