Gryps: Institutional-Grade RFQ Perpetuals

The largest pools of perpetuals flow stay off-chain because the venues on-chain were not built for them. Order books expose size before a fill. Automated market makers concede slippage that scales with the trade. Centralized exchanges solve execution at the cost of custody. For institutional desks, protocol risk managers, and any participant whose orders move markets when they hit a public book, those tradeoffs make on-chain perpetuals a marginal venue rather than a primary one.

Gryps is the protocol being built to close that gap. It is an institutional liquidity layer for intent-based RFQ perpetuals on SEI, designed around private trade intent, firm solver quotes, zero-slippage execution once a quote is accepted, and bilateral non-custodial settlement.

What Gryps is

Gryps is an execution layer. The taker is a fund, trader, a desk, a treasury, a vault, or a protocol acting on behalf of its users. The supply side consists of professional solvers who quote on demand, hedge externally, and compete for fills. Solvers source liquidity and pricing across the major centralized and decentralized perpetuals venues. Binance is fully integrated today. Hyperliquid is next. The relationship with these venues is partnership rather than competition. Gryps brings their depth into on-chain non-custodial execution rather than asking institutional flow to pick between the deepest off-chain books and the right custody posture.

Settlement is bilateral and non-custodial. Gryps v1 was built on Symmio’s clearing layer, and is currently transitioning to its own internal Orbs smart contracts as additional venue integrations come online. Gryps is the routing, quoting, settlement, and operating system that ties these together.

Gryps is market-structure infrastructure for the flow that public order books, AMM pools, and CEX-shaped venues do not serve well at scale: large block trades, private bilateral arrangements, long-tail asset exposure, and programmatic protocol hedging. Gryps was built for serious DeFi participants to route through when size, privacy, pricing, and settlement integrity matter, rather than another front end competing for retail order-book volume.

Why existing venues struggle for this flow

A visible order book is a public broadcast of directional information. For a trader moving size, or a protocol hedging an exposure that is itself observable on-chain, the act of placing the order is a signal that other participants will trade against. The cost shows up as worse fills, predictable price reactions, and execution outcomes systematically below expectations.

Automated market makers handle continuous flow well. Compression at size is a separate problem. The deeper an order, the worse the average fill, and the more aggressive the post-trade reversion that other participants harvest. Many on-chain venues also require liquidity to sit in advance, paid for with emissions, hoping the right flow arrives. The capital is least available exactly when it is most needed, and can be expensive to maintain.

Centralized exchanges remain the default for size because they solve execution. The cost is custody, jurisdictional exposure, counterparty concentration, and operational friction that institutional teams now actively want to reduce. Going non-custodial on-chain has historically meant accepting thin depth and meaningful slippage in exchange. Gryps is solving for this tradeoff.

How an RFQ trade works on Gryps

The mechanism is an intent-based RFQ.

Submit intent. The taker defines market, side, and size. The intent is private. It is not posted to a public book and is not visible to the broader market.

Route to eligible solvers. Gryps routes the intent to solvers eligible to quote that asset, size, and risk profile. Eligibility reflects asset coverage, quote history, and risk limits.

Receive firm quotes. Solvers price the trade using their internal models, inventory, and external liquidity sourced across the integrated venues, then return executable quotes for the requested size and a defined time window. Firm means the quote is a price the solver is committing to, not an indication.

Accept the best quote. The taker accepts the best executable price. There is no mechanical price-impact curve to fight and no public book to sweep.

Settle bilaterally and non-custodially. Settlement creates an isolated risk relationship between the taker and the winning solver. Risk is per-counterparty, not pooled across the venue.

Solver hedges externally. Having taken on the position, the solver lays it off using its multi-venue inventory and CEX depth. Gryps does not need to own all the liquidity it executes against,it needs to be the venue through which that liquidity becomes reachable on-chain.

The properties this lifecycle delivers are the ones the flow has been waiting for. Directional intent is never broadcast, so privacy holds before fill. Solvers commit to the prices they quote, so once a quote is accepted, execution occurs at the quoted price. There is no curve to compress the trade, so size is tolerated. Any solver capable of pricing an asset can quote it without Gryps committing fixed market-making cost per pair, so long-tail coverage extends naturally. And settlement does not require handing custody to a centralized counterparty, so asset control persists until execution.

Bilateral, non-custodial settlement

The settlement layer carries as much weight as the routing layer in this design. Many institutional concerns about on-chain venues are concerns about pooled risk: shared liquidity that absorbs losses, custodial intermediaries that sit between counterparties, and generalized exposure that creates contagion paths in stressed conditions. Bilateral settlement isolates risk per counterparty relationship. A taker has exposure to its specific solver, not to a generalized pool. Assets remain under user control until execution.

Version 1 of Gryps was built on Symmio’s clearing layer, with a 12-hour fraud-proof window protecting against disputed clearings before USDC is released. In Version 2, Gryps is currently transitioning to its own internal Orbs smart contracts as additional venue integrations come online. Owning settlement is the architectural prerequisite for running multi-venue solver coverage properly. Per-counterparty risk isolation, programmatic margin and collateral handling, and the operational surface institutional integrations require all run through this layer.

Who Gryps is for

Gryps is built for institutional trading desks, funds, and treasuries managing positions that move when they hit a public book, and for professional traders who care about execution quality at size. The constraints that decide whether they route flow through Gryps are direct: privacy of directional intent, firmness of quoted price, settlement integrity, depth on the assets they need, and operational standards that meet what their internal counterparties expect.

Solvers and DeFi-protocol integrations are part of how the venue works rather than who it sells to. Solvers price the flow and make the venue more useful to the customers above.

How Gryps is positioned in the market

Order-book venues handle visible, retail-shape flow well. AMM pools are strong for continuous on-chain liquidity. Centralized exchanges offer deep size and fast settlement, with custody as the cost. Each venue type is genuinely useful for the flow it was designed to serve.

Gryps is designed for a different execution problem. Private, firm-quote, block-size perpetual execution for size-sensitive flow. The defensibility of the position is architectural rather than promotional. RFQ plus bilateral non-custodial settlement is not a feature an order-book venue can adopt without dismantling the model that makes it competitive at retail scale. Solver relationships compound over time. Protocol integrations accumulate. The institutional posture and operational standards that the design requires are not a register a retail-first venue can credibly switch into.

The wedge is narrow on purpose. RFQ is not universally better than every other primitive. For institutional-grade flow on-chain and traders who value the same qualities it serves, it is the right one.

The Gryphon, briefly

The Gryphon stands for a synthesis. The lion carries the discipline of institutional execution: precision, accountability, and the standards a market earns by being held to them. The eagle carries the sovereignty of DeFi: composability, freedom from custody, and programmable settlement. The protocol exists where those two halves combine. The aim is that institutions, traders, and serious DeFi participants do not need to compromise on either half to use it.

The final point

Order books, DEX’s, and centralized exchanges each serve a real role in this market and will continue to, and Gryps is partnering with them. The flow Gryps is built for sits not exactly within their natural shape. Private intent, multi-venue solver competition, firm quotes, bilateral non-custodial settlement, SEI-native execution: these are what the Gryps is being built around.

Companion piece: Trading on Gryps: The Execution Workflow covers what the venue looks like in operation.