🐝Voodoo V2 - HIVE

Introducing HIVE

Hybrid Intent and Vault Engine, or HIVE for short, is a novel implementation that combines the speed and UX improvements of intents with the advantages of decentralised, onchain p2p liquidity pools.

How do existing intent solutions work?

UniswapX, Cowswap, 1inch Fusion and Symmio are RFQ (request for quote) systems where users trade tokens using intents.

Instead of directly calling a smart contract to trade, users define an end goal, for example "swap 1 ETH for 2500 USDC" or "long 1 ETH at 5x leverage". This end goal is executed by "solvers" who are incentivized with a fee to achieve this end state. Solvers can be bots, professional market makers or protocol run servers.

While the former are spot platforms, Symmio has pioneered an RFQ system for perpetual derivatives. IntentX, Alpha Thena etc. are 3rd party projects which provide a UI for Symmio. The relation is akin to SNX and Kwenta, where SNX is the smart contract system and Kwenta is the UI.

RFQ systems maintain an off-chain, centralized orderbook akin to Ethereum's mempool where orders are queued for execution. Professional market makers execute these orders if they find it profitable.

Intent Pros

Executions are gasless, you simply sign a message and the trade is performed. You don't have to pay a network fee in ETH. However, solvers recoup their fee by charging a percentage fee on your tokens.

High speed: Intent-based exchanges can give CEX-grade speeds since the orderbook is off-chain. Though in practice, UniswapX and 1inch take a couple of minutes on average for tokens to settle.

Paired with account abstraction (ERC 4337), Intent-based projects can provide CEX grade UX with social login functionality, removing the need of managing keys on Metamask.

Cons

While the current crop of intent-based exchanges took one step forward for trading speed and UX, they took two steps backward for decentralization and the interests of retail traders.

Firstly, they give a lot of power to professional market makers, to the detriment of retail users. GMX has a peer to peer pool model, where the pool acts as a counterparty to user positions. If a trader longs 1 ETH, the pool has an opposing position of 1 ETH short. Retail LPs simply need to put passive liquidity into the GLP pool, without having to worry about market making bots and algo trading strategies. In the current form, Symio / IntentX remove LPing and vaults from the equation altogether. The only way to profit is to trade against professional market makers, where the odds are stacked against retail. You cannot deposit liquidity in a pool in the hope of yield. The relationship between traders and LPs is done away with in this instance.

If the role of MMs is to simply fill retail positions, what advantage do they have? Unlike the blockchain, "intent-based mempools" are opaque. There is a real possibility of MMs colluding to salami slice profits from slippage levels defined in intents. Slippage loss and frontrunning happens in DEXes, but the data is visible on-chain. Traders are then subjected to the same opacity of CEX order books, masked in the guise of DeFi.

GMX LPs earn via a number of avenues, not just perps trading. A significant portion of earnings come from spot trading and liquidations. These earnings are not part of the equation with intent-based projects.

Intents in the current form kill composability. The USP of GMX was access to real yield, made possible only by passive LPing in GLP vaults. Thanks to vaults, GMX was able to offer escrowed GMX tokenomics that reward long term LPs. 3rd party yield optimizers like GMD and Beefy would not be possible without this composability. Retail traders are losing these yield opportunities with current intent-based exchanges.

Voodoo V1

Voodoo has its own implementation of intents, before the term became popular. To open or close margin positions, traders send a **request transaction**. If slippage and price constraints are met, the voodoo keeper server opens the positions.

The keeper server is the heartbeat of Voodoo, responsible for opening positions, liquidations and updating the oracle price. In the current Intent lingo, this is the "searcher".

The keeper maintains an off-chain orderbook for limit orders. These are executed when their limit price is met. However, unlike RFQ protocols with opaque orderbooks, the Voodoo orderbook is completely on-chain. Matching happens off-chain but state storage and settlement is on-chain. The smart contract has a flag [`Vault.inPrivateLiquidationMode`](https://github.com/voodoo-trade/voodoo-contracts/blob/094e4f7d69c9f0ef3187645bd73d197e846ef4d2/contracts/core/Vault.sol#L77) to enable 3rd party searchers to compete for liquidations in return of a fee.

Voodoo V2 Hybrid Intent and Vault Engine (HIVE)

The Voodoo team has watched the intent landscape evolve and is thrilled to announce HIVE. HIVE is a next-gen perp AMM combining intents with composable LP vaults. HIVE combines the advances made in UX by intents with the security and fairness of peer-to-peer LP pools.

Hybrid pool + RFQ model: Retail investors simply need to LP into VLP pool. The protocol will use part of the funds for market making on a protocol manged, builder-maintained orderbook. This is an orderbook with CEX level speeds. The idea is that traders who want speed trade on the RFQ market, while those opting for security of the blockchain use the smart contracts. Using a common pool increases capital efficiency as liquidity is not fragmented. LPs earn yield from both avenues.

Gasless transactions: HIVE utilizes ERC-2612 based gasless approvals, bringing the experience in-line with intent-based systems like Cowswap and perp DEX infra providers like Symio.

Tailor made intent-pool based on MPC: HIVE has its own intent propagation infra based on multi-party computation (MPC) for security and decentralization. This retains value with the protocol's investors instead of leaking revenue to rent-extracting third party infra providers.

Frictionless upgrade: Existing VLP pool assets will be moved to v2 , with no action needed on users' end. This ensures a seemless migration experience for Voodoo LPs.

Retain the non-synthetic model for security while listing more pairs: GMX's token borrowing model is guaranteed to be solvent; it needs no insurance pool. Insurance-based protocols like Mango have been bankrupted in the past. At the same time, Voodoo will list top performing tokens to offer traders more variety. Comparing GMX v1 (borrowing) to v2 (synths), version 1 is still the dominant protocol by TVL, volume and users.