A Copy Agent mirrors one wallet’s Meteora positions automatically: when they open, we open; when they exit, we exit. The wallet you chose to copy is the leader. It is one of Alokumi’s two Agents; the other is the Strategy Agent, which runs your own LP rules (see What rules can express).
What it does
- Every 30 seconds it reads the leader’s positions, works out how much we should hold in each pool in the leader’s proportions, and moves your holdings toward that target.
- Exits always run before any new open. Exits are checked every minute; when the leader leaves a pool, we follow on the next check.
- New opens and top-ups happen every 5 minutes, at most half of each pool’s gap at a time (at least $25), so a new pool usually fills in two rounds; at most 6 new pools per round on paper funds, and one open or top-up per round with real funds.
- It follows every pool the leader holds; there’s no fixed limit on how many.
- Ranges follow the leader’s. When they split one range into several neighbouring positions, we open across their combined range.
- Fully automatic: no Watch mode, no semi-automatic confirmations. It calls no AI model and doesn’t use your daily AI quota.
Every open, top-up and exit goes through the same safety checks as a Strategy Agent (pool and token checks, the sell-back test, per-position and total budget limits). If a check fails, nothing happens and the ledger shows Not done: … naming the check.
Picking a leader
There are three ways in: New Copy Agent on the Agents page, and Copy this wallet on each leaderboard row and on a wallet’s page.
Step 1, Pick a leader, lists the latest smart-money leaderboard with a tick or a cross for each check. Only wallets that pass every check can be copied:
| Check | Meaning |
|---|---|
| On the leaderboard | On the latest leaderboard, and qualified |
| ≥ 20 closed positions in 90 days | Enough history |
| Profitable in both USD and SOL (90 days) | Realized PnL positive in both units |
| Active in the last 14 days | Still trading |
Missing data counts as a fail. The first 10 are shown; Show {n} more lists the rest.
Budget and limits
Step 2, Budget & limits:
| Setting | Meaning | Default |
|---|---|---|
| Copy budget (paper funds) | Spread across the leader’s pools in the same proportions as their open positions. At least $50 | $1,000 |
| Per-pool cap | Most we put in any one pool. Empty = no cap | 40% of budget |
| Total deployed cap | Most in positions at once; the rest stays as cash | 90% of budget |
| Stop-loss line (% of budget) | If equity falls this far below the budget, we close everything and stop copying. Empty = none | 30% |
| Minimum order | Smaller differences are skipped (and shown in the ledger) | $10 |
| When the leader exits | Exit all at once: fastest; exits always run before any new open. Exit in slices: each round we exit only as much as the pool can absorb at about 1.5% price impact, to soften the impact on thin pools | All at once |
| Max slippage (under Advanced) | Applies to every swap the Agent makes: buying in, range resets and selling back after exits. 0.5%–5% | 1% |
How the shares work: shares follow the leader’s open positions, not their whole wallet. When the leader keeps cash aside, we still deploy up to your caps. Example: the leader holds $6,000 in pool A and $4,000 in pool B; your budget is $1,000, total deployed cap $900, per-pool cap $400. Proportionally that’s A $600 and B $400; A is first trimmed to the $400 per-pool cap; together that’s $800, under the $900 total cap, so nothing else shrinks (above it, every pool scales down by the same factor). Whatever a cap trims is shown on its own ledger line, naming the cap.
When equity is below the budget, targets follow equity, not the budget: losses don’t get topped back up.
Step 3, Start on paper funds. The name is optional; the default is “Copy” plus the leader’s short address. You can pause it or close all positions any time on the Agent page.
Reading the ledger
On the Agent page, each Copy ledger row is one moment in one pool: what the leader did → what we did, then “After: target / held / gap”. For example:
Leader opened $5,000 → We opened $200 (in steps) After: target $400 · held $200 · gap $200 (50%)
Positions the leader already had when copying started show as Leader holds, not “opened”. Skips are logged with their reason too: below minimum order, cap reached, open limit reached this round, leader data stale, leader no longer qualifies.
The four numbers at the top:
| Metric | Meaning |
|---|---|
| Capture rate | While copying, how much of what we should hold (in the leader’s proportions) we actually held, time-weighted; the build-up period doesn’t count |
| Exit delay | Seconds from the leader’s exit to ours. Two columns: seen → done, on-chain → done |
| Build-up time | From the start of copying until holdings reach 95% of target (watched for up to 30 minutes) |
| Over-allocation | The part we held above target; not counted in capture rate |
Time with leader data unavailable, and time after the leader stops qualifying, don’t count toward capture rate. The ledger tab also has a table of every exit’s delay and each pool’s target (hover to see the amount before your caps).
When it stops
| Situation | What happens | Alert |
|---|---|---|
| The leader no longer qualifies | Exits only: no new opens or adds. If the leader later reduces a pool we hold, we reduce in proportion | Leader no longer qualifies. Pick another leader or stop |
| Leader data stale (over 3 minutes) | No action that round | None |
| The leader loses 8% or more in one UTC day (open + realized) | No new opens or adds until 00:00 UTC; exits continue | Leader down {n}% today, plus a banner at the top of the Agent page |
| A pool’s copy capacity is full (real funds only) | All real-funds Copy Agents together may hold only what the pool can absorb if everyone exits at once (about 1.5% price impact), first come first served. An open or add is trimmed to fit, or skipped; exits are never limited. Paper funds aren’t limited | A ledger line Pool copy capacity full; the Agent page and the leader step show X: Pool copy capacity full |
| A pool is too shallow for real-funds copying | Even with no Copy Agent in it, the pool can absorb less than one minimum order within that price impact, so no real-funds Copy Agent ever opens there. Waiting doesn’t help; paper funds aren’t limited | A ledger line Pool too shallow — real-funds copying skips it; the Agent page and the leader step show X: Pool too shallow — real-funds copying skips it |
| Leader data down for over 15 minutes | Following exits only, no new opens; resumes automatically when data is back | Leader data down {n} min, plus a banner at the top of the Agent page |
| Leader data incomplete (positions without amounts) | Exits only | A ledger line |
| The stop-loss line is hit | Everything is exited and copying stops | Stop-loss hit, copying stopped. Start a new Copy Agent to continue |
| You pause it | Only exits and the stop-loss line run; no new opens | — |
| You close all positions | Every position is exited; the leader is no longer followed | — |
Alerts show up in To do on Home and in Telegram.
Paper funds and real funds
Paper funds first: no real money moves, but prices, fees and swap costs follow the real market. Paper costs come from the same sources as real funds: real Jupiter quotes for swaps, and fees from your share of the bins the price actually crossed. See how closely it tracks (capture rate, exit delay), then decide about real funds.
Real funds for Copy Agents are open to some accounts only for now. If yours is one of them, step 2 offers Real funds (shadow first):
- Shadow mode first: the Agent plans every move and asks the Policy, but sends nothing. What it would hold shows in the copy ledger and the Deployed figure. Real execution starts once the platform switches it on.
- Funds sit in the Agent’s own on-chain Vault, the same as a Strategy Agent; see Use real funds. Real-funds Copy Agents have a budget cap per Agent and per account for now; the create page says when you’re over it.
- Swap between SOL and USDC (on by default): if the Vault holds only SOL and the leader’s pool is quoted in USDC (or the other way round), the Agent swaps first and swaps back to the token you deposited after the exit. Each swap costs slippage and fees. When it’s off, those pools are skipped and you’re told which token to deposit.
For the general difference between paper and real funds, see Use real funds and Security.