A Hyperliquid whale-tracking and copy-trading bot. On-chain wallet intelligence, edge-scored whales, slippage-guarded mirroring — every copy sized and risk-checked before a single order is placed. Hyperliquid today, multi-exchange execution built in.
Hyperliquid is fully on-chain, so every whale's position, fill, and equity is public and readable in real time — no API key, no permission. WhaleEdgeBot turns that raw feed into a ranked, scored, copyable shortlist.
Subscribe to a curated watchlist or auto-discover from the Hyperliquid leaderboard. Opens, scale-ins, partial closes, and flips are detected the moment they hit the chain — via WebSocket, not block-polling.
A wallet isn't a whale just because it's big. Minimum account value, track record length, win rate, drawdown ceiling, and leverage-style filters strip out one-lucky-trade wallets and un-mirrorable degens before scoring even runs.
Each qualified whale gets a 0–100 edge score from ROI, win rate, profit factor, a Sharpe-like measure, and drawdown. Scores gate copying, scale position size, and auto-pause a whale whose live edge decays.
Set auto_discovery.enabled: true and the leaderboard scanner sources fresh candidates on a schedule, runs them through the same filters and scoring, and surfaces only those clearing your threshold. A decaying score auto-pauses copying — no whale rides your account down on reputation alone.
A whale's $2M position can't be copied one-to-one. The engine translates their Hyperliquid trade into a sized, leverage-capped, slippage-guarded order on your exchange — or skips it cleanly if it can't be mirrored faithfully.
| Whale | Trade | Lev | Edge | Action |
|---|---|---|---|---|
| whale-A | BTC Long | 14× | 91 | COPY |
| whale-C | ETH Long | 8× | 87 | COPY |
| whale-B | SOL Short | 5× | 79 | COPY |
| whale-F | HYPE Long | 3× | 74 | COPY |
| whale-D | kPEPE Long | 10× | 58 | SKIP |
| HL coin | Venue sym | Sizing | Cap | State |
|---|---|---|---|---|
| BTC | BTCUSDT | 5% eq | 10× | PLACED |
| ETH | ETHUSDT | 5% eq | 8× | PLACED |
| SOL | SOLUSDT | 5% eq | 5× | PLACED |
| kPEPE | 1000PEPEUSDT | — | — | LOW EDGE |
| FARTCOIN | not listed | — | — | UNAVAIL |
The infrastructure under the copy logic. Two-layer venue separation, slippage and latency guards, per-whale allocation, drift reconciliation, and a live dashboard — one Rust core, nothing bolted on.
Tracking is locked to Hyperliquid on-chain reads. Execution is your pick of twelve venues. Watch whales on Hyperliquid, mirror onto Binance.
# tracking ≠ execution
Tracking: hyperliquid
Execution:
ExchangeName: binance
Fixed notional, % of your equity, proportional, or mirror-leverage. The whale's size becomes yours — scaled by edge score and per-whale weight.
mode: percent_of_my_equity equity_percent: 5.0 apply_edge_score: true
If price has run past the whale's fill or detection took too long, the copy aborts. A late entry chasing a moved market is the fastest way to lose the edge.
max_slippage_percent: 0.5 max_entry_delay_seconds: 15 order_type: market
No single whale dominates your book. Max notional, open copies, daily loss limit, and consecutive-loss auto-pause — all enforced per wallet.
max_notional_per_whale: 500 pause_after_losses: 3 daily_loss_limit: 100
The hard part isn't entering — it's exiting. Full and partial closes mirror automatically, with an independent stop in case you miss the whale's exit.
copy_full_close: true copy_partial_close: true independent_stop: 8%
Every 30s the engine re-checks your book against each whale's true state and corrects missed partials or orphaned positions before drift compounds.
reconcile_interval: 30s auto_correct: true close_orphans: true
PostgreSQL for production, SQLite for local — same codebase, driver-detected at startup. Per-whale PnL attribution computed from day one.
Driver: sqlite | postgres
Metrics: per-whale PnL
win rate · drawdown
Whale opens, copy fills, skips, and risk events on separate channels. Every card shows whale label, edge score, leverage, and resolved venue symbol.
#web-whale-activity #web-copy-fills #web-risk-alerts
Three independent layers — per-copy, per-whale, and account-wide — run concurrently on every position. Any layer can flatten a trade. A whale's leverage is never trusted blindly.
Each mirrored position carries its own stop, firing even if the whale holds. Because you're always a step behind their fill, you never rely solely on their exit to protect you.
Exposure to any single wallet is capped by notional and equity percentage. A losing streak auto-pauses the whale; a daily loss limit benches it until the next session.
An account drawdown ceiling halts all new copies and can flatten the book. A margin buffer refuses copies that would follow a whale toward liquidation.
Signal-only mode runs the full tracking and copy engine against real whales — detecting, scoring, sizing, and logging exactly what it would place, without sending a single order. Watch the would-copy log build conviction before you risk a dollar.
Run place_order: false, dry_run: true. Review the would-copy log: confirm symbol resolution, sizing, and which whales clear your edge gate against real activity.
Flip to place_order: true with one whale, minimum size, isolated margin. Compare real fills and slippage against the dry-run estimates. Tune your edge gate and caps.
Add whales and size from your own per-whale PnL attribution — objective data, not a curated highlight reel. Drawdown and win rate justify each step up.
WhaleEdgeBot doesn't hammer Hyperliquid alone — RedisMarketBridge does the heavy reads once, for the whole bot family. Whale state, fills, and pre-scored edge data land in Redis. The bot reads structured opportunity data and executes.
Signal-only tracking is always free. Pay only when you switch on live copying. License key delivered instantly — the bot goes live on the next restart, no config rewrite, no reinstall.
WhaleEdgeBot v1.0 — signal-only tracking free forever. Switch on copying when the data tells you which whales have earned it. No credit card to start, no lock-in if you continue.