3.4 seconds flat.
Three banks.
35,000 customers.
Cross-border settlement didn’t need to take 5 days. We co-built the payment rail that proves it — giving bank customers a dedicated on-chain address and bidirectional access to stablecoin rails, live in production with three financial institutions.
on the rail
vs 5 days
per year
Cross-border money still moved on 1970s infrastructure.
An American fintech was building the product that would let bank customers move money between traditional accounts and the blockchain. Three banks had signed on. The technology stack didn’t exist.
The correspondent-banking network is elegant in 1972. By 2024 it means your $10,000 transfer sits in a queue for five business days, gets inspected by intermediaries, converted twice, and arrives with $30 in fees stripped out. Stablecoin rails settle in under four seconds, around the clock, with no correspondent in the loop.
One rail. Every user gets their own address.
The core architectural decision: instead of a shared pool, every bank customer receives a dedicated on-chain address. Deposits and withdrawals are unambiguous by design. No reconciliation table. No manual matching.
account
on-chain address
settlement
account
on this rail today
At scale, the rail carries 35,000+ dedicated addresses simultaneously — every customer’s settlement path is always unambiguous, never shared.
Open source. No shared pools. No vendor lock-in.
Every architectural decision was made with one constraint: the rail must work at scale without introducing fragility. We were part of the core engineering team and co-owned these choices.
One address per user. Always.
The first question in any payment rail is reconciliation: when money arrives, whose is it? Most platforms solve this with a shared pool and a matching layer that can fail. We solve it architecturally: every bank customer gets a dedicated on-chain address at account creation. When funds arrive at that address, attribution is unambiguous — no matching table, no reconciliation run, no edge case.
Bidirectional. Bank account in, bank account out.
The rail moves money in both directions: bank account to blockchain and blockchain back to bank account. Customers hold USDC or RLUSD on-chain or elect to have the platform settle instantly back to their traditional account. The legacy rail remains an option — the platform handles routing. We built and operated the bridge layer that made both paths work without customer-visible complexity.
Open-source core. No proprietary lock-in.
The payment-rails platform was built open source under the hood, built to order. That means the client owns the codebase, can audit every line, and is never at the mercy of a licensing change. We believe financial infrastructure should run on open protocols — the same principle behind Pix, behind HTTP, behind every standard that has outlasted the vendors who tried to replace it.
Scale from 5,000 to 20,000 users — no rearchitecture.
The three banks came with very different sizes: 5,000 customers, 10,000, and 20,000. The platform handled all three without a different deployment configuration. Volume scales linearly — address provisioning is cheap, the settlement layer is shared, and the platform’s stateful components are partitioned to avoid contention. We also co-built the operational tooling so the fintech team could monitor and triage across all three banks from a single interface.
The numbers from a live financial platform.
These figures are from a running production system, not a pilot. The rail has been settling transactions since launch — no babysitter, no scheduled downtime, no SLA breach.
the same rail
customers
vs 5-day legacy
per year
Is your settlement
process still running on
1972 infrastructure?
Whether you’re building a new payment product, adding stablecoin rails to an existing banking app, or solving a cross-border settlement problem nobody has cracked cleanly yet — tell us what you’re trying to settle.
No sales team. Mathias reads every message.