Hyperliquid vaults
Vaults are Hyperliquid's native way to follow a strategy with pooled capital: you deposit, a vault leader trades the pool, and your share moves with the pool's equity. Mirroring is the opposite arrangement. Nothing is pooled, every position exists under your own address, and your exit does not depend on anyone else's.
In short
A Hyperliquid vault pools depositors' capital into a single account run by one vault leader; your exposure is a share of that pool, positions are collective, and you are exposed to that leader alone. HyperMirror reproduces leaders' positions inside your own account — up to 10 scored leaders in Full mode, one isolated sub-account each — so attribution, margin and exit stay individually yours.
| Dimension | Hyperliquid vault | HyperMirror |
|---|---|---|
| Capital location | Deposited into the vault contract | Your own account and sub-accounts |
| Exposure | Vault leader's single strategy | Score-weighted basket of leaders |
| Attribution | Pooled vault PnL | Per-leader PnL in isolated sub-accounts |
| Exit | Vault withdrawal rules and lockups | Revoke agent, positions close, funds already yours |
| Fee model | Vault profit share | 0.1% builder fee on mirrored volume |
HyperMirror fits when you want several independent edges, your own custody, and the ability to close positions yourself at any moment.
Hyperliquid Vaults may fit better when you have specific conviction in one vault leader and prefer the operational simplicity of a single deposit.
A vault is not custodial in the exchange sense — it is a protocol construct, not a company holding your balance — but your capital does leave your own account and enter a shared one. That changes what you control. You are a pro-rata claimant on a pool, subject to the vault's lock-up and withdrawal rules, and you cannot close an individual position because no individual position is yours.
Mirroring keeps the boundary intact. Capital stays in your Hyperliquid account and sub-accounts; the agent may place orders and nothing else. If you disagree with a position at 3am, you can close it from your own wallet without asking anything of the operator or waiting for a redemption window.
Vault leaders are usually compensated with a profit share on the pool's gains. That aligns the leader with upside and costs nothing in a losing period, but it also introduces the classic option-like incentive: the leader captures a slice of the upside without symmetric exposure to the downside, which can encourage variance.
HyperMirror's 0.1% builder fee applies to mirrored notional volume and nothing else — no profit share, no subscription, no spread markup. The trade-off is stated plainly: a high-turnover leader generates more fee at the same profitability than a low-turnover one, so fee cost scales with activity rather than with results.
In a vault, execution happens once for everyone. The leader submits an order, the pool takes the position, and every depositor's share moves identically. There is no slippage difference between depositors and no tracking error, which is genuinely elegant.
In a mirror, execution happens per user. Your account receives its own orders sized to your capital, which means your fills can differ slightly from the leader's — that difference is tracking error, and it is the price of holding your own positions. In exchange you get per-leader isolation: leader A's long in one sub-account cannot be netted away by leader B's short in another, so both edges stay expressed and each leader's contribution stays measurable.
A vault's risk profile is the vault leader's risk profile, and depositors inherit it wholesale. If the leader takes a position that goes badly, the drawdown is the pool's drawdown, and your only lever is redemption under whatever terms the vault sets.
A scored basket dilutes that. A single leader's collapse moves a weighted fraction of the portfolio rather than all of it, soft-issue strikes and replacement act on measured decline rather than on your own reaction time, and emergency removal exists for events that should not wait for a scheduled review. None of that protects against a correlated market move that hits every leader at once.
Choose HyperMirror if: you want several independent decision-makers rather than one, you want funds to stay in your own account with no lock-up, and you want per-leader attribution rather than a pooled equity curve.
Choose Hyperliquid Vaults if: you have researched one vault leader and want exactly that exposure, you prefer a single deposit with no per-user execution differences, or you actively want the leader compensated on profit rather than volume.
Most people asking "vault or mirror?" are really asking six smaller questions. Answer them in order and the choice usually makes itself, because the two structures differ on every one of them.
Choose a vault if you have genuine conviction in one specific vault leader, you are comfortable with a pro-rata claim on a shared pool, your capital is small enough that a single deposit is the only capital-efficient option, and you actively prefer the leader to be paid on profit rather than on volume.
Choose a score-weighted mirror if you would rather not stake the outcome on one decision-maker, you want funds to stay in your own account with no lock-up, you want per-leader attribution and per-leader isolation, and you want selection and replacement to follow a published rule set instead of your own reaction time.
Choose both if you want to: they are not mutually exclusive. A vault deposit and a mirrored basket are separate exposures, and running them side by side is a legitimate way to compare a pooled single edge against a diversified isolated one with real money and honest bookkeeping.
Both are leveraged perpetual-futures exposure on one venue. A correlated market move can hit a vault and every sleeve of a basket in the same hour, diversification across leaders is not diversification across market regimes, and no structure converts a bad month into a good one. Structure decides how a loss is distributed, attributed and exited — not whether losses happen.
Where HyperMirror is limited: isolated sub-accounts fragment margin and are less capital-efficient than a pooled account; per-user execution introduces tracking error against the leader's fills; and Starter mode means one leader until the $100k unlock.
Where Hyperliquid Vaults is limited: your outcome depends entirely on one leader; positions are not individually yours and cannot be closed selectively; and withdrawal timing is governed by the vault's terms rather than by you.
HyperMirror's side of every row above follows one published rule set. The Elite basket is sticky: a leader stays mirrored until a rule removes them, and a higher-scoring wallet elsewhere never forces a swap. Emergencies — account value below roughly $1,000, or no fill for 96 hours or more combined with zero trades in 7 days — remove a leader immediately. Everything softer accrues at most one strike per wallet per UTC calendar day, with three strike-days triggering replacement and a clean day resetting the counter; drawdown for that test is measured on jump-adjusted equity so deposits and withdrawals are not read as trading losses, and ROI is measured from PnL. The full table is on How it works and in the documentation.
Structure changes which risks you carry, not whether you carry risk. Nothing on this page is a return estimate, a recommendation, or a claim about another provider's results — check any provider's own documentation before deciding.
Past performance is not indicative of future results. Perpetual futures are leveraged instruments and carry a substantial risk of loss, including the loss of your entire position.
Methodology
Scoring and replacement are documented in full on How it works and in the Docs (Policy v3). In short: the Elite basket is sticky, emergencies remove a leader immediately, and soft issues accrue at most one strike per UTC day with three strike-days triggering replacement. Read how it works or the documentation for the full table.
Questions
Not in the exchange sense — it is an on-chain construct rather than a company balance sheet — but your capital does move into a shared account and your claim becomes a share of the pool rather than specific positions.
Yes. The positions are in your own account and sub-accounts, so you can close them from your wallet at any time and revoke the agent.
You can, and that is a reasonable way to diversify. You would then be doing the selection, weighting and replacement work yourself, which is the part a scored basket automates.
No. Fills differ because your orders are your own; that divergence is tracking error and it can work slightly for or against you.
Keep reading
Score-weighted allocation across up to 10 elite Hyperliquid traders, each isolated in its own sub-account. Your funds never leave your account.
Non-custodial · Agent cannot withdraw · Cancel delegation anytime
