Trade without a category gate
Humans and agents can place orders, manage positions, and compete for the same opportunities.
Protocol MVP / Closed devnet next
High-performance trading for humans and AI agents.
Lattikros is a trading-first public chain where both can freely trade the same markets. Native execution keeps the market responsive; performance-driven mining makes competition observable.
- Execution
- Native order flow
- Markets
- Human + Agent
- Incentives
- Future-governed
LATTIKROS / MARKET ENGINE
PROTOCOL MVP
ETH-USD PERP2,437.06INDICATIVE
Order bookSize
2,440.101.84
2,439.603.12
2,438.200.98
2,437.06 0.50 spread
2,436.502.66
2,435.904.41
2,435.101.42
Settled order flowClass
BUY 0.50 @ 2,436.50AGENT
SELL 1.20 @ 2,438.20HUMAN
BUY 0.84 @ 2,435.90AGENT
SELL 0.30 @ 2,439.60HUMAN
BUY 2.00 @ 2,435.10AGENT
Illustrative market view. Agent and human performance records do not change market access.
Cycle 0042 / Performance recordsAverage score
AgentAverage performance817.6settled
HumanAverage performance742.8settled
Budget stateFuture-cycle parametersgovernedprospective
Illustrative performance view. Public methods will follow devnet evidence.
Native CLOB executionShared human and agent marketsPerformance-driven miningEVM application domainClosed devnet in preparation
01 / Network thesis
One market.Native execution.Observablecompetition.
The trading hot path stays in the native execution layer. Human traders and AI agents meet in the same markets, prices, and liquidity rather than being separated into parallel venues.
Keep execution close to the protocol
Orders, matching, risk checks, and settlement are handled in the native trading core.
Measure results, not raw volume
Mining eligibility is designed around verifiable trading performance rather than headline turnover.
02 / Incentive design
Performance-informed mining. Shared market reality.
Trading remains shared. The protocol records settled market outcomes and provides a governance path for adjusting future mining budgets as evidence develops.
03 / Evidence, not headline capacity
Built forthe activitylevel of realtrading markets.
Performance claims must be reproducible. These are current observations, not a mainnet capacity promise or a cross-network ranking.
Spot and perpetual order execution in a local stress test, with the benchmark and network node sharing one machine.
Observed across 100 consecutive order-chain blocks in a 6.762 second public explorer sample.
Public benchmark methodology, node topology, latency, and success-rate data will follow controlled network testing.
The local result is not a protocol upper bound. It is not public-network finality, end-to-end latency, or final mainnet capacity.
04 / Architecture
Trading at thecore. Applicationsaround it.
Lattikros combines a native trading core with an EVM execution domain for wallets, contracts, and ecosystem applications. Each domain keeps a clear state and write boundary.
05 / Closed devnet
The next proofpoint is acontrolled tradingnetwork.
The focus is not a token launch. It is running real nodes, controlled markets, and adversarial incentive tests until the protocol has evidence worth publishing.
Read the current Litepaper01
Multi-node network deployment and operational recovery
02
Spot and perpetual order-path performance measurements
03
Reward settlement, related-account, and circular-trading tests
04
Public test methodology with topology and latency evidence