Latency
<10
ms, end to end
An orderbook exchange on zk rollups: CEX speed, your keys, settled to Ethereum.
A central limit orderbook running offchain for speed, with every state change proven to Ethereum by a zero-knowledge validity proof. Swap against the same depth. One price, either way you trade.
Latency
<10
ms, end to end
Throughput
200K+
orders per second
Settlement
ZK
validity proofs to Ethereum L1
Crypto, equities, FX and commodities as perpetual futures: cross or isolated margin, trailing stops, partial close, and gasless orders. Long or short, the account is still yours.
Leverage amplifies losses as well as gains.
Say it in plain language, and it becomes a working bot: explicit rules you can read, a backtest you can question, a deployment you can stop. No code. It trades around the clock, on spot or perpetuals, and the account it trades from never stops being yours.
A representative view of the LifyX terminal. Example figures, not live market data.
Buy ETH whenever it falls three percent below its twenty-day average, sell it back at the average, and stop out if it drops another six.
Spot, perpetuals and automation draw on the same balance and report to the same state. Nothing moves between venues, because there is only one.
Orders are signed from your own wallet, an MPC social login or an EOA you already hold. The exchange never takes custody, and the proofs on Ethereum are the record of that.
In your own wallet: an MPC social login or an EOA you already hold. Every order is signed from it, the exchange never takes custody, and settlement is proven to Ethereum rather than promised.
The orderbook runs offchain, which is where the sub-10 ms latency comes from. Zero-knowledge validity proofs then commit every state change to Ethereum L1, so the chain verifies the math, not the operator.
Spot against orderbook depth, swap against the same book, and perpetual futures on crypto, US equities, FX and commodities: 177+ markets, up to 100× leverage, cross or isolated margin. Leverage amplifies losses as well as gains.
You describe a strategy in plain language; it becomes explicit rules you can read. Backtest over past data first, then deploy on spot or perpetuals. It runs from your account and you can stop it at any time.