A parallel EVM, so your contract calls don't queue behind everyone else's. Portfolio margin that nets a hedged book to one number, not three. Matching that doesn't bleed to MEV. The things every serious trading desk has been asking for, on a chain that ships them.
Most new exchanges compete on lower fees. We're not interested in that race. MetaFlux is a bet that traders care more about market structure, risk infrastructure, and execution latency than they care about saving 0.2 bps a fill.
RFQ for size that shouldn't advertise on the book. Frequent batch auctions for markets where fairness beats latency. Matching encrypted at intake, revealed at settlement.
A SPAN-style scenario engine nets your hedged book to one number. A T0 yellow-card grace tier freezes resting orders before any position is touched — one block to react, not one tick.
Transactions execute optimistically and in parallel; conflicts re-run, the rest commit. Derivatives-aware precompiles (portfolio margin, ADL pricing, RFQ depth, vault NAV) run native — not as 80,000-gas Solidity.
Any builder can list a new perpetual or spot market via an on-chain gas auction. No protocol-team gate, no allow-list, no committee. The protocol arbitrates the auction; you ship the market.
Win the gas auction, supply the spec (leverage, maintenance ratio, funding parameters, oracle source), and the market is live next block. Liquidity is your problem; the protocol does the rest.
MIP-3 spec →Same address format, same signing, same Solidity. Plus derivatives-aware precompiles (portfolio margin, ADL pricing, RFQ depth, vault NAV) callable from any smart contract.
EVM docs →The gateway speaks the wire formats your stack already speaks. Same URL shapes, same request/response, same EIP-712 envelope. Point your client at MetaFlux, change one chain ID, keep trading.
Approve a trading key once from your cold master. The agent signs every order without ever holding withdrawal authority. Per-bot scoping, automatic expiry, a clean audit trail.
Agent wallets →POST /info and POST /exchange with the wire shapes your stack already speaks. WS subscriptions with the channel names you already subscribe to. Your existing fill handlers and order placement code keep working.
Migration guide →For deeper coverage, the documentation at docs.mtf.exchange has the complete API surface, concept catalog, and integration guides. This is just the highlights.
An independent L1 derivatives chain. Its own validator set, its own BFT consensus, its own EVM execution layer. Built from first principles for derivatives clearing, not borrowed from a general-purpose L1.
Specifically: Hyperliquid, dYdX (v4), Lighter, Aster. The honest table:
| Capability | MetaFlux | Hyperliquid | dYdX v4 | Lighter | Aster |
|---|---|---|---|---|---|
| L1 BFT consensus | yes | yes | yes | L2 | yes |
| EVM-compatible | yes | partial (HyperEVM) | no (Cosmos) | no | partial |
| Parallel EVM execution | yes | no | n/a | n/a | no |
| Cross-asset portfolio margin | yes | limited | no | no | no |
| T0 grace tier before partial | yes | no | no | no | no |
| RFQ + FBA matching | yes | no | no | no | no |
| MEV-resistant matching | yes | no | no | no | no |
| Permissionless market deploy | yes | yes | no | no | no |
| Drop-in HL-wire compat | yes | native | no | no | no |
Capability is one axis. Liquidity, fees, listings, and operational maturity are others. Pick the chain that wins on the axis you actually care about.
Mainnet follows audit and testnet milestones, not calendar dates. The roadmap targets a multi-month testnet phase with phased validator onboarding. Subscribe to updates via the docs site.
If your bot already speaks a wire-compatible protocol, change one URL and one chain ID. For greenfield builds, the MTF-native SDK comes in TypeScript (with WASM-backed crypto for browser performance) and Rust. Either way, the EIP-712 envelope is standard and every EVM wallet works unmodified.
Genesis validator recruitment opens ahead of mainnet. The initial set is small and grows through the first year. Reach out via the validator inquiry below — we want operators with on-chain ops experience and geographic diversity.
USDC via Circle CCTP, natively attested. Other assets via a single audited third-party bridge, finalised pre-mainnet via a public ADR. We deliberately did not build a monolithic in-house bridge: blast radius reduction.
The blue / pink / white reads cleanly as a brand system and isn't taken by anyone else in this space. It also signals what kind of community we want: technical, welcoming, and uninterested in the dominant aesthetic of the rest of crypto.
Testnet access is open for market makers, builders, and integrators. We're particularly interested in feedback from teams shipping production trading systems.