01SERA-PROPRIETARY
Virtual Liquidity
One deposit quotes against every pair. No per-pair pre-funding. Capital never sits idle.
Sera protocol · v2.1 · mainnet
The settlement engine stablecoins were missing.
A four-plane system engineered for cross-border stablecoin FX: Virtual Liquidity, single-transaction Smart Order Routing, gasless intent execution, and non-custodial settlement audited by CertiK. Built for the corridors that clear trillions in real economic flow, not speculation.
Quote latency
300ms
median · p50
Settlement
≈1 blk
on-chain finality
Stablecoins
600+
supported assets
Pairs quoted
30k+
synthetic + direct
Capital eff.
>100×
vs. pooled AMMs
Audit
100%
CertiK · live monitor
The proprietary FX engine.
Not an AMM retrofit.
FX doesn't need price discovery — it needs execution, compliance, and capital efficiency. Two world-firsts power the stack: Virtual Liquidity and single-transaction Smart Order Routing. Neither can be replicated by an AMM, DEX, or legacy settlement network.
02SERA-PROPRIETARY
Smart Order Routing
Best execution across the full book. Generalised flow with gain factors. All-or-nothing settlement.
03GASLESS · MEV-RESISTANT
Intent execution
Users sign one EIP-712 intent. Relayers assemble. Sandwich attacks collapse at the batch boundary.
04AUDITED · NON-CUSTODIAL
Security model
Funds never leave user wallets until on-chain settlement clears. Formally specified in TLA+. Continuous monitoring.
§ smart order routing · in practice
An exotic pair with no direct liquidity. Sera finds the route anyway.
Across 30,000+ pair combinations, only a fraction have direct stablecoin liquidity. For everything else — MYR → BRL, IDR → NGN, KRW → ZAR — the router stitches a path through the most liquid intermediates and settles it atomically in a single binding transaction.
Direct attempt — no liquidity existsMulti-leg execution — 3-hop atomic settlement
332,000 lines of production code. None of it forked.
Sera's V1 codebase has been independently measured using COCOMO and SLOC — the same methodology private-equity firms apply during technical due diligence. The result: 332,557 lines of production code across 1,925 files, with a cyclomatic complexity score of 20,477.
Replicating this stack from scratch would cost an estimated US$12.01 million in organic development and require 35+ months of concentrated effort by a 30-engineer team.
Lines of code
332,557
production · 1,925 files
COCOMO cost
$12.01M
organic · replicating V1
Schedule effort
35.4mo
concentrated build
Complexity
20,477
cyclomatic · system-wide
Complexity at this scale is not boilerplate. It is the algorithmic density of an order-matching engine, a settlement state machine, and a multi-venue routing layer compiled into one execution stack.
Virtual Liquidity 01 · Sera-proprietary
One stablecoin. Quotes across every pair. Capital that is never idle.
Most AMMs lock LPs into a single pool — your USDC sits in USDC/EURC, deaf to the ten other corridors that need it. Sera indexes every LP position into a unified balance sheet, then synthesises quotes against every supported pair on demand. A single USDC deposit can back swaps in BRLA→MXNB, XSGD→IDRX, and EURC→JPYC simultaneously.
- Single-asset deposit, multi-pair earnings across every active corridor
- Greater than 100× capital efficiency versus pooled AMMs
- Auto-rebalanced against on-chain price feeds with deviation monitoring
- Reference band aggregated daily from 50+ central banks, 10+ PSPs, and 10+ FX houses — defines what "off-market" means
§ the fork problem
Why generic AMM primitives fail at FX.
Every existing on-chain FX attempt is built on a fork of a generic AMM — Uniswap-style constant-product (x·y=k), Balancer-style weighted pools, or Curve-style stableswap invariant. They aren't poorly built. They're the wrong primitive for the job. None were designed for FX, and they all share one fatal flaw for a basis-point market: asymmetric price discovery. Pricing is fair in only one direction of the curve; the reverse blows out.
Asymmetry is just the headline. The deeper failures are structural. AMMs are single-leg only — they can't atomically route SGD → USDC → KRW, and FX without multi-hop is half a product. They're LP-driven, not market-maker driven — passive depth profiles can't service institutional flow. They're capital-inefficient at scale, needing 10×+ TVL to match the spreads of an off-chain venue. And slippage scales with size, which makes them unworkable for the million-dollar tickets institutional FX runs on.
They also force every cross-currency trade through a USD intermediary, paying the spread twice.
Sera is a purpose-built CLOB, not an AMM. Two-sided pricing. Market-maker depth. Multi-leg atomic routing — direct N×N pairs without a USD bottleneck. Virtual Liquidity makes it ~100× capital-efficient against any pooled architecture.
Global FX is $9.6 trillion a day. Forking an AMM won't service it.
Topologysame trade · two protocolsBRLA → IDRX
Generic AMM
BRLA → USDC → IDRX
47 bps
Sera CLOB
BRLA → IDRX
12 bps
Asymmetric pricingsame pool · same size · two directionsEUR ⇌ JPY · 1M
Forward→EUR-stable to JPY-stable1,000,000 trade
Reverse←JPY-stable to EUR-stable1,000,000 trade
Generic AMM
8.4× wider one way
major non-USD cross · curve still asymmetric
Sera CLOB
1.0× equal both ways
two-sided book · symmetric quote
Four ways to settle FX. One that actually works.
Side-by-side against the alternatives: generic AMMs forking Uniswap or Curve, CEX-style remittance rails like Wise, and crypto-native CLOB DEXes built for derivatives, not currency.
| Sera · CLOB | AMM fork | CEX / bank rail | CLOB DEX (derivs) | |
|---|---|---|---|---|
| Price discovery | Two-sided MM book | Bonding curve · asymmetric | Internalised | Two-sided · derivs-tuned |
| Multi-leg atomicity | Native · single tx | Single-leg only | Off-chain reconciliation | Not supported |
| Capital efficiency | >100× · Virtual Liquidity | 1× · per-pair pooled | High · but pre-funded | Margin-based |
| Settlement custody | Non-custodial · ~1 block | Non-custodial | Custodial · T+1 to T+3 | Non-custodial · synthetic |
| MEV exposure | Intent-batched · resistant | Sandwich-vulnerable | N/A · off-chain | Sequencer-mitigated |
| USD bottleneck | Direct N×N pairs | Forces USD hop | Correspondent banking | USD-quoted only |
| Pairs supported | 30k+ synthetic + direct | Per-pool | ~40 corridors | ~50 perps |
| Cost at $1M | Zero slippage · binding quote | +108 bps reverse · slippage | Hidden in spread | Order-book dependent |
| Audit · third party | CertiK · TLA+ spec | Varies | Regulator-attested | Varies |
Smart Order Routing 02 · Sera-proprietary
Single-transaction path-finding across direct, multi-hop, and synthetic routes.
For exotic corridors with no direct pair (say BRLA→IDRT), the router stitches the trade through the most liquid intermediates, settles it in a single transaction, and presents one binding quote. No slippage drift. No failed routes. No manual hop selection. The whole route lives or dies in one transaction.
- Direct, multi-hop, and cross-stablecoin synthetic paths in one quote pass
- All-or-nothing execution — partial fills are architecturally impossible
- Generalised network flow with gain factors, sub-200ms median quote latency
Route finder · evaluating direct vs. routedBRLA → IDRT
chosenBRLA → USDT → IDRT187ms · 0.22% · 1 tx
Gasless · MEV-resistant 03
Users never see a gas dialog. Searchers never see your trade.
Sera abstracts gas at the protocol level. Fees are deducted in the swapped stablecoin, paid by the relayer, and reconciled out-of-band. Users sign one EIP-712 intent; the relayer assembles the transaction. Order intents are batched and revealed only at execution, which collapses the sandwich-attack surface. For institutional flow, Sera supports private mempool routing through a permissioned relayer set.
- Pay gas in any supported stablecoin — no native token needed
- Intent-based design eliminates sandwich attack surface
- Private mempool routing for institutional order sizes
01
User signs EIP-712
pay X · get Y · min-out · deadline
0 gas
↓
02
Relayer batches + routes
SOR · private mempool · MEV-shielded
relayer pays
↓
03
On-chain settlement
atomic · all-or-nothing · ~1 block
paid in USDC
Four planes.
One protocol.
End-to-end determinism.
Every swap on Sera moves through four separable planes. Each is independently upgradeable, independently monitored, and independently auditable. Below: the live data flow from user signature to on-chain settlement.
01 · Intent layer
02 · Routing engine
03 · Liquidity plane
04 · Settlement plane
Intent
Users sign an EIP-712 intent — pay X, get Y, with min-out and deadline.
single signatureno gas · no approvals
In · user wallet
Routing
Off-chain SOR explores direct, multi-hop, and synthetic paths. Quote binding for 30s.
<200ms path-find40+ venues
Liquidity
Virtual Liquidity engine maps LP balances to active quotes across every pair on demand.
30k+ pairs quoted>100× capital eff.
Settlement
All-or-nothing atomic settlement. The whole route lands or none of it does.
~1 block finalityCertiK-audited
Out · local FX
Engineered to clear trillions, audited like it.
01
Formal specification
Every contract path specified in TLA+ before implementation. Invariants property-tested in CI on every commit.
02
Independent audit
CertiK-audited across the full contract surface. Differential audits on every protocol upgrade. Reports public.
03
Continuous monitoring
Multi-source reference band — 50+ central banks, 10+ PSPs, 10+ FX houses. Tracks oracle deviation, LP skew, mint/burn anomalies, gas-price irregularities. On-call paged within 30 seconds.
04
Dev community
Bug bounty live on Immunefi. Collaborative audits. Researchers contribute continuously to the security posture.
The settlement engine is a proof, not a promise.
\* Settlement.tla — invariant: atomic all-or-nothing
\* Property: no partial fills, no orphan transfers
MODULE Settlement
EXTENDS Naturals, Sequences
VARIABLES routes, executed, refunded
Atomicity ==
\A r \in routes:
\/ r \in executed
\/ r \in refunded
\* Never both. Never neither.
SpecOK == [](Atomicity /\ NoDoubleSpend)- Spec → CIEvery contract path is specified in TLA+ before implementation. Invariants are property-tested on every commit.
- AuditCertiK-audited across the full contract surface. Differential audits on every protocol upgrade. Reports public.
- MonitorMulti-source reference band tracks oracle deviation, LP skew, mint/burn anomalies, gas irregularities. On-call paged within 30 seconds.
Dutch auction AMM. Coming soon.
V2 extends the protocol with a Dutch auction AMM purpose-built for foreign exchange. Details on release.
V2 · classified until release
Build on the settlement layer.
Fintechs and stablecoin issuers ship cross-border products on Sera without the banking overhead.
Fintechs & neobanks
Ship a cross-border product in weeks, not quarters.
Sera's API handles pay-in, FX swap, and pay-out in a single call. No license stack to assemble, no liquidity to source, no settlement risk.
Read the docs →Stablecoin issuers
Bring real utility — and liquidity — to your stablecoin.
List on Sera's book to route through every remittance and B2B flow the protocol carries. Non-USD stablecoins get immediate international utility.
Talk to BD →