Custodial exchange copy trading versus a non-custodial agent

What custody actually adds

A custodial structure introduces obligations that have nothing to do with trading: the operator must stay solvent, stay operational, process withdrawals promptly, and keep your assets segregated from their own. Each of those is a separate way to lose money while the strategy itself is performing fine.

That does not make custodial platforms illegitimate. Regulated, well-capitalised custodians manage this competently every day. It means you are underwriting an additional risk and should price it.

What the agent model changes

Hyperliquid separates trading permission from transfer permission at the protocol level. An approved agent can submit order actions; it cannot sign a withdrawal, cannot move funds to another address and cannot change account ownership.

• Deposits go from you to your own Hyperliquid account — no platform address in the path.
• Positions, balances and history stay auditable on-chain under your own address.
• You revoke the agent from your wallet without needing the operator's cooperation.
• The 0.1% builder fee is capped by a separate approval that grants no trading or transfer rights.

The honest trade-offs of self-custody

Self-custody moves key management onto you. Lose your wallet and no support desk can restore it. Custodial platforms also net margin across a single account and can offer simpler onboarding for people who do not want to hold keys at all.

Isolated sub-accounts, which prevent leaders' positions from netting, deliberately give up some margin efficiency in exchange for clean per-leader attribution.

Risk statement

Removing custody risk does not remove trading risk. A trade-only agent can still open leveraged positions that lose money or get liquidated.

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.