Conductor

The keeper layer for v4 hooks. Your hook reacts, Conductor acts, your fees land where you said. Enforced on chain. CA: 0xdd215018a9c558fcdfb2db56ab4d8c60ddf25491 useconductor.xyz/

Late question for the builders still up. What's one thing you'd let an AI agent suggest for your onchain setup, but never do on its own?

Found an automation recipe you like on Conductor? Fork it. You fill in your own inputs, they're checked against the recipe's bounds, and you sign the create from your own wallet. Your automation then shows which recipe it came from, and the recipe's fork count goes up.

How much can an operator charge for a run? The smallest of five numbers. What the run actually cost plus margin, the max per run, the task's own cap, the fee budget for the period, and what the automation can afford. If the fee gets capped, the run still goes through. The operator just gets less.

Every run on Conductor gets a one line explanation in plain English, labelled "Written from on-chain facts". The sentence is built from the run's own data first. If the rewording adds a number, an address or an action that isn't in that data, it's thrown out.

Want your Conductor alerts in a Discord channel? The docs have a small receiver you can copy. It checks each alert's signature with our SDK, then posts the summary and a link to your channel. If it answers with an error, Conductor retries the delivery.

Builder rewards on Conductor have a holder gate, and the details matter. To be paid you hold at least 0.05% of $CONDUCTOR supply, measured as a time weighted average over the whole round, across your wallet and up to 2 linked ones. A balance held for one block barely moves that average. And holding alone pays nothing. Only real users of what you built do.

Conductor's v4 pool checker reads a pool's price and liquidity straight from the v4 StateView, and can wait for one of five things: Any change at all Price leaves a range Price is back inside a range Liquidity drops below a level Liquidity rises above a level Any automation can use it as its trigger today.

Every suggestion an AI agent makes on Conductor has a status you can follow. It only counts as applied when a verified receipt matches the transactions you signed. If you dismiss it, the same suggestion from that key is muted for 7 days. If nobody signs, it expires after 72 hours. The agent suggests. Your wallet decides.

How an agent key works on Conductor. It's shown to you once and stored only as a hash. It covers 1 to 20 of your automations, expires after 7, 30 or 90 days, and you can revoke it in one click. There is no scope for withdrawing funds, changing ownership, adding a recipient or unpausing. Those actions don't exist for agents.

You can describe a Conductor automation in one sentence. Here's what happens to that sentence. Our model turns it into a draft in a closed format. It never writes calldata. Plain code then compiles the draft into the same policy and task the step by step builder makes. Every row of the plan says where its number came from: your own words (underlined in your sentence), an answer you gave, something we looked up, or a Conductor default. Every limit says who holds it. Either the contract rejects anything over it, or it's an operator rule with the chain caps still behind it. Then the plan is test-run against the current chain state, and your wallet signs it in the normal create flow. Plans written in plain English are capped at 1 ETH a day. Writing the draft doesn't need a wallet at all. useconductor.xyz/app/new

Four things the Conductor policy engine does on every run that are easy to miss. An operator can only fill the blanks you left. Every other byte of each call has to match your template exactly. It never uses delegatecall. It measures what actually moved, before and after each call, instead of trusting what the call says. Token approvals are reset to zero before the run ends.

Which keeper job for v4 hooks should we ship first? Rebalance a position when the price leaves its range. Refresh a hook's oracle between swaps. Step a dynamic fee with volume. Run a large order in slices. Top up a pool when its liquidity runs thin. Reply with yours. The most-asked one goes first.

Most keepers check your pool on a timer. Conductor can wake up the moment something happens. An automation can watch a v4 pool's price and liquidity, and the worker re-checks it as soon as a swap in that pool is indexed, instead of waiting for the next interval. You can set it up today from a blank automation. Ready-made templates for these jobs are coming.

Two Hook Studio blocks that work together: max swap and launch guard. Max swap refuses any single swap above a share of the pool. Launch guard adds a tighter cap for the first hours or days after the pool opens. When both apply, the lower cap wins. Snipers hit a wall right at launch, and normal traders never notice.

Every Uniswap v4 hook on Robinhood Chain has its own page on Conductor. It shows the permissions read from its address, every pool it runs, each pool's live tick and fee, the last swap, and which keeper jobs fit it. The biggest hook right now runs 11,819 pools and did 3,600 swaps in the last day. useconductor.xyz/hooks

How does an operator become verified on Conductor? It needs 50+ runs at 98%+ success in open mode. Then the registry owner can grant it, and can take it away again. Verified is what new automations allow by default, so it has to be earned on chain, not asked for.

New on Conductor: Hook Studio Market v2 You can now sell the v4 hooks you build. Set a price per launch, earn up to 0.5% of every swap on every pool launched from your hook, or sell the whole listing to someone else. Conductor does not take a cut. useconductor.xyz/market