Live reference monitor
Pyth ETH/USD
$-,---.--+0.00 bps session
Session startLatest signed price
PublishedWaiting
Freshness-
Confidence-
Samples0 / 36
Server-authenticated Pyth Hermes feed. No wallet or simulated value is used here.
Reactive liveFallback verified
UHI10 Sustainable liquidity × MEV protection
Fees should follow outcomes.
Not fear.
MARKOUT is a Uniswap v4 hook that escrows a provisional charge, waits for post-trade evidence, then lets an authenticated observation return it to good flow - or retain it when LPs were adversely selected. The Fair-Flow profile starts at 18 bps, while Reactive Network supplies a live event-to-action rail.
TRADE_QUALITY_ENGINE DETERMINISTIC
SWAP
PRICE
REACTIVE
POST-TRADEMARKOUTt + 5 min
EXECUTION→OBSERVATION→FAIR FEE
768seeded tradessix market regimes
$1.999Mvolume per policyidentical trade tape
+21.87%vs fixed feeLP net after proxy
0medium / highSlither findings
Judge proof mode
Four public proof paths. No wallet required.
Select an outcome, inspect its measured state, then open every source, settlement, and claim directly in a public explorer. The evidence boundary stays visible beside the result.
Economic lifecycle verified
18 bps complete rebate
A real Uniswap v4 swap matured, received a Pyth observation, finalized at the base fee, and executed the sponsored claim path.
18 bpsfinal fee
100%surcharge rebated
55sfallback relay
Evidence boundaryThis is the deployed 18 + 50 bps Fair-Flow profile and its full-rebate terminal branch.
Market intelligence layer
Watch the price. Then watch the fee respond.
The live panel reports signed Pyth ETH/USD observations. The replay panel uses declared deterministic paths to expose both settlement extremes without pretending that a controlled experiment is a live trade.
LIVE MARKET DATAPyth-signed ETH/USD observations CONTROLLED REPLAYFrozen synthetic mechanism study LIVE TESTNETTransactions remain in the console below
CONNECTING
Controlled scenario replay
Five-minute outcome window
ScenarioNear execution
Directional markout0.0 bps
Protocol clock0:00 / 5:00
Execution · t+0Pyth observation · t+5m
Base fee18.00 bps
Trader rebatePending
LP protectionPending
Final feePending
OutcomeWaiting for delayed evidence
Live testnet console
Make the swap. Watch the fee change.
This is the deployed USDC/WETH Uniswap v4 pool on Unichain Sepolia - not a simulation. Your wallet signs every transaction; no private key enters this page. Reactive Network watches the canonical evidence event, while an independent fallback keeps the live demonstration recoverable.
Unichain Sepolia · 1301
Gas0 ETH
Spendable0 USDC
Spendable0 WETH
Pool pending0 USDC
LP reserve0 USDC
01
Execute
Real v4 swap
USDC
18 bps pool fee+50 bps provisional
Approval appears first only when the selected token allowance is insufficient.02
Observe
Onchain lifecycle
1
Swap minedWaiting
2
5-minute markoutWaiting
3
Pyth evidence publishedWaiting
4
Reactive subscription eligibleWaiting
5
Fee finalizedWaiting
Execute a swap or load an existing trade to start the live timeline.
03
Verify
Fee allocation
Final effective feePendingNo active trade
Escrowed-
Returned-
LP protection-
Directional markout-
Active trade ID
No trade selectedTestnet only · contracts are experimental and unaudited · Pyth ETH/USD assumes test USDC ≈ $1.
Frontend proof surface
The interface explains the trade. The contracts decide it.
MARKOUT turns a delayed cross-chain lifecycle into a judge-readable receipt without moving pricing, settlement, or custody into the browser.
Show the 18 bps base, the refundable 50 bps charge, and the maximum possible fee before the wallet signs.
The browser wallet approves only when needed, then submits the swap directly to the Fair-Flow testnet pool.
The interface reads maturity, publication, delivery, and terminal allocation from chain state; it does not decide the result.
The final fee, rebate, LP reserve, claim action, and explorer receipts remain visible as one auditable settlement receipt.
Connect→Preview→Swap→Observe→ClaimEvery terminal number is read from the deployed hook.
Guided comparison
Same market. Different information.
Live Fair-Flow pool: 18 bps base + refundable 50 bps surcharge. The console and public lifecycle above use the separately deployed Fair-Flow hook; original 30 + 50 bps evidence remains independently linked.
Average effective fee by flow class
Benign flow
Directional markout≈ 0 bps
The trader did not capture a meaningful post-trade move.
Fixed feeSame for every trader
30 bpsVolatility feePrices the market regime
49.48 bpsMARKOUTPrices the realized outcome
27.43 bpsOutcome
Fair-Flow returns most of the provisional charge and finishes below the ordinary 30 bps pool.
Reactive-driven mechanism
Swap → mature → react → allocate
Reactive Legacy callback publicly verifiedMarkoutRequested→Reactive maturity→Authenticated callback→Allocation finalized
Separated system architecture
Four planes. One outcome-priced hook.
Each plane has one job. The frontend explains, Unichain accounts, Ethereum verifies, and Reactive Network turns the verified event into an authenticated action. A separately authenticated fallback preserves liveness.
01
Experience planeFrontend
Wallet console
Preview, swap, monitor, claim.
Evidence surface
Fee receipt + explorer links.
02
Economic plane · UnichainUniswap v4 + MARKOUT
beforeSwap / afterSwap
18 bps base + bounded provisional escrow.
MarkoutHook
Immutable trade, directional curve, rebate and LP reserve.
SettlementCoordinator
Authenticated first-valid delivery; duplicates are no-ops.
03
Evidence plane · EthereumPyth + publisher
Signed price update
Price, time, confidence and market checks.
Canonical event
One normalized observation for one trade.
Independent fallback
Attested redundancy with four public relays.
04
Reaction plane · Reactive NetworkEvent-to-action rail
Exact subscription
Pinned publisher, event signature and market.
ReactVM execution
Decode the normalized observation and encode the destination action.
Authenticated callback
Live autonomous cross-chain reactionCallback proxy + injected RVM identity.
Wallet swap→MarkoutRequested→Pyth observation→Reactive action→Hook allocation
GOOD FLOWRebate to trader
ADVERSE FLOWLP protection reserve
NO DELIVERYPermissionless full refund
Reactive Network integration
Reactive makes verified evidence actionable.
Sponsor thesisLive event-to-action rail
The Legacy RSC pins Ethereum Sepolia, the publisher contract, event signature, and exact market topic.
One canonical Pyth publisher event activates the stateless reaction and no MARKOUT-operated listener is required.
The destination receiver requires both the system callback proxy and the injected RVM identity.
The coordinator forwards the first valid delivery once; duplicates no-op and expiry preserves the full-refund guarantee.
Why it materially matters
The hook cannot hear Ethereum. Reactive bridges evidence into action.
MARKOUT's hook lives on Unichain, while the canonical Pyth-verified observation is published on Ethereum Sepolia. The Legacy RSC subscribes to that exact event, runs in ReactVM, and requests the authenticated destination callback. Reactive Network therefore provides the primary autonomous cross-chain execution path. A separately authenticated fallback enters the same replay-safe coordinator without taking pricing authority.
Exact event subscription✓Authenticated callback✓Custody×Fee authority×
Controlled mechanism study
A frozen dataset, declared constraints, and a result that can fail.
The primary result is a reproducible controlled synthetic study where each policy receives the same tape. A separate mainnet replay below tests the fee ordering on canonical swap events. Both use a pool-level adverse-selection proxy rather than exact LVR or individual LP profit.
Dataset · markout-phase-6-v1
768 trades · $1.999M per policySix regimes: benign, informed, inventory-improving, mixed low-vol, mixed high-vol, and stale/manipulated reference.
SplitMix64 seed · 2026082501Freeze one tapeIdentical trades for all policies
→02Run three policiesFixed · volatility · MARKOUT
→03Sweep 21 base fees10 through 30 bps
→04Select by constraints18 bps is first eligible
27.4262 vs 30 bps at equal execution.
Full surcharge rebate; 18 bps final.
Modeled net-after-proxy versus fixed.
Historical robustness replay · June 1, 2024
251 real mainnet swaps. Directional discrimination survived.
Canonical Uniswap v3 USDC/WETH events were replayed with a five-minute same-pool reference. MARKOUT charged favorable outcomes least and adverse outcomes most. Aggregate LP net did not beat fixed in this short window, and that negative result remains visible.
251eligible real swaps
$3.188Mobserved notional
5 minmarkout horizon
-0.39%aggregate LP net vs fixed
Full provisional rebate.
Slightly below fixed on average.
More LP protection retained.
What survived: the intended 18.00 < 29.02 < 39.14 bps directional ordering on observed events. What did not generalize: the synthetic study's +21.87% aggregate LP-net result; this window measured -0.39% versus fixed.
Public testnet proof
Four lifecycles. Both settlement extremes.The independent fallback delivered Pyth observations for four real Unichain v4 swaps. These receipts prove that MARKOUT remains recoverable around the primary Reactive Network rail: both terminal extremes, the Fair-Flow 18 bps rebate path, and the sponsored-claim entrypoint are publicly verifiable.
4 public end-to-end lifecycles3 / 1 full rebates / full retention38/67/67/55s measured relays
LP net after proxy · USDC
$3,889
Fixed fee30 bps avg fee
$7,594
Volatility48.53 bps avg fee
$4,740
MARKOUT34.25 bps avg fee
Trader adoption proof
Good flow wins without assuming deeper liquidity.
Under the selected 18 bps base, benign and inventory-improving traders already beat a fixed 30 bps pool at equal execution quality. Informed flow pays more only after its realized outcome is adverse.
+$2.57saved vs fixed per $10k
+$22.05saved vs volatility per $10k
Wins the fee-only route at equal execution quality.
+$12.00saved vs fixed per $10k
+$29.44saved vs volatility per $10k
Receives the full surcharge rebate and pays only the 18 bps base.
−$31.06trader saving vs fixed per $10k
−$13.02trader saving vs volatility per $10k
Intentionally expensive: adverse flow funds LP protection.
Measured result: benign flow saves $2.57 and inventory-improving flow saves $12.00 per $10,000 versus fixed, while MARKOUT still improves modeled LP net-after-proxy by 21.87%. Boundary: routing, demand elasticity, and additional liquidity remain unmodeled.