The Playbook

What AlphaWhisper is, and isn't

What it is: a tool that surfaces where institutional dealer hedging will pull, pin, or amplify price, then scores when that structure aligns with smart-money flow and intraday tape into a real setup, across three trading horizons:

• 0DTE day trading, Gravity Intraday + Orbit on 1m/5m/15m timeframes, KING/STORM/FLIP read for same-day expiry, Hawks flow alerts ≤15s old, the per-(strike × expiry) gamma map for SPX/SPY/QQQ.
• Intraday swings, Quick Swing engine (1-5 day horizon) scoring confluence across Gravity + Hawks + Darkpool + Index Sync. Setups fire on the dashboard and Discord when the score crosses 8/9.
• Multi-week swings, Multi-Week engine (5-20 day horizon) requiring persistent Hawks flow, heavy darkpool accumulation, and ≥3:1 R:R. Lower volume, higher conviction.

The signals are mechanical, not opinionated. Same engine, same rules, every symbol, every session.

What it isn't: an autopilot, a guaranteed win-rate, or a replacement for risk management. Every signal is a hypothesis, not a buy button. The House Rules at the bottom of this page exist to keep you alive long enough for the math to work.

Best for: option traders who want a structural read on SPY/SPX/QQQ for intraday + 0DTE scalps, and ~100 liquid single names scanned daily for swing setups. Works whether you trade once a week or twenty times a day.

Your first 5 minutes

If you do nothing else, do these in order:

1. Open the dashboard and read the Track Record section
   Tells you the current edge per confluence-score bucket (5/9 vs 8/9) and which of the 6 criteria actually drives wins. This is the honest read of whether the engine is paying.
2. Check the 🎯 Quick Swing + 📅 Multi-Week cards on the dashboard
   Compact cards showing today's high-conviction setups. Tap any card for the full detail modal (entry, target, stop, R:R, the reasons it fires).
3. Open 🧲 Gravity for SPX/SPY/QQQ
   The dealer-positioning map. Use the expiry pills to pick your timeframe (0DTE for day-trading, weekly for swing). KING/STORM/FLIP at the top are the key levels for the session.
4. Open 🪐 Orbit to see price action with GEX bands overlaid
   TradingView-style price chart with dealer-magnet bands drawn on the candles. Use timeframe pills (1m/5m/15m/1D) for your trading horizon. Fullscreen button top-right.
5. Join the Discord channel so you get the alerts when you're away from the screen
   Hawks alerts every 15s (whales printing), Quick Swing alerts 3×/day at 9:30/12:00/15:45 ET, Multi-Week ~2-3×/week. All are filtered for high-conviction and freshness, no spam.

The app tabs explained

The top nav (desktop) or bottom tab bar (mobile) has five views. Each owns a different question.

📊

Dashboard

Home / triage

What it shows: Market Pulse (the whole platform's read in one strip), the Focus Trade of the Week (the single highest-conviction idea, deduped against the quick-swing list), the Big Move Radar (movers with a BULLISH / BEARISH bias badge and a 🌀 coiled-spring tag for names wound up to move), today's confluence setups, and the Verdicts for tracked symbols.

How to use it: First stop every morning. Verdicts come in three tiers: full setups (all conditions met), Leaning (a directional tilt, size small, collapsible detail), and Wait. Cards are sorted by score; click any symbol to drill into the detail panel.

🧲

AlphaWhisper Gravity

Dealer-positioning map

What it shows: The Readout strip (GRIP, KING, FLOOR, CEILING, FLIP, ± MOVE, the pre-entry numbers at a glance), then the strike ladder with GEX magnitudes and darkpool prints. Three sub-tabs:

• Intraday, the ladder you trade off, with the GEX / DEX / VEX / CHEX lens pills and the FOCUS / FULL FIELD depth switch (tight decision band vs every tracked strike).
• Horizon, multi-date matrix of dealer stacking. Weekly mode = 5 Fridays for swings; Daily mode (indices) = 0DTE plus ~2 weeks of sessions, with each column's own day-KING ⭐, OPEX chips, expected-move and week-over-week badges.
• Whisper, the live 0DTE agent: fired setup cards with lifecycle stamps, the WATCHING board, and the Track Record strip. Full guide in section 7.

How to use it: Read GRIP first (FIRM = fade edges, LOOSE = respect momentum), then KING / STORM / FLIP as the session's structural anchors. Trade with-the-magnet (long FLOOR / short CEILING) or watch for breaks of FLIP for regime shifts.

🦅

AlphaWhisper Hawks

Institutional flow tracker

What it shows: Active "whale" bets, single-contract trades ≥$500k premium with ≥75% on the ask side (real buyers, not mixed flow). Each whale tracked from print → current mid → P&L since fill.

How to use it: Look for tickers where Hawks bias aligns with the Verdict and Gravity direction. A whale's coattail is much more useful when the structural read agrees. Filters: Fresh-only, by direction, by premium.

🪐

AlphaWhisper Orbit

Price chart with overlays

What it shows: A native TradingView-style candlestick chart with GEX strike-bands drawn behind the candles, KING/STORM/FLIP labeled lines projecting forward, and darkpool prints as purple dots anchored to the candle they hit.

How to use it: Pick a timeframe (1m/5m/15m for day trade, 1D for swing). The floating OHLC tooltip top-left follows your crosshair. Click ⛶ top-right for fullscreen. The chart refreshes live every 15-30s, your zoom stays put.

Mantis

LIVE · in testing

What it shows: a desk blotter of Gravity-gated swing setups on single names and the index ETFs, each with a full plan: entry, a volatility-scaled stop with a close-based invalidation, two structural targets with their R:R and roles, 1R in dollars, the management plan, what the plan pays on the named contract, the structural read, and the other side of the trade.

How it decides: dealer positioning is the gate (no dealer thesis, no card), a balanced book is withheld, and every card is graded against real prices and real premium. See section 10b below.

💼

Portfolio

Track your trades vs the engine

What it shows: Your open + closed option positions with live mid-price and P&L, scored against the engine's signal at entry.

How to use it: Add positions manually as you enter them. Over time you'll see your hit rate on engine-aligned vs engine-contrary trades, that's the feedback loop that makes you better.

The lexicon

Every term in the app, defined once. Bookmark this section.

Gravity levels (Strike-zone vocabulary)

Flow + microstructure

The Readout strip + map depth

Reading the Gravity map

Open Gravity, Intraday and you get a price ladder: one row per strike, biggest premium at the widest bar. Every row reads left to right the same way, and every badge on it has a precise meaning. This is the full legend, one flashcard per symbol with an example.

Zone pills, what kind of level it is

The colored word next to the value tells you how dealers behave at that strike.

Special markers, the badges after the value

These call out the most important strikes on the map.

The two swing engines

Both engines score setups out of 9 points using the same 6 confluence criteria, calibrated for different time horizons.

The 6 scoring criteria (max 9 pts)

1. L1, Setup zone (+1 or +2): spot near FLOOR (+2) or MAGNET (+1)
2. L2, Target structure (+2): KING above spot in actionable range
3. L3, Regime (+1): spot above FLIP (positive-gamma mean-revert mode)
4. L4, Smart-money lean (+1 or +2): Hawks bullish bias ≥2 (+2) or ≥1 (+1)
5. L5, Institutional positioning (+1): darkpool accumulation at/below spot ≥$50M
6. L6, Cross-tape (+1): Index Sync · Bullish

Mirror logic for SHORT setups (CEILING / STORM / below FLIP / bearish Hawks / DP distribution / Index Sync bearish).

The Whisper agent 🦉

Whisper is the deterministic 0DTE decider that rides the live Gravity loop on SPX / SPY / QQQ. It is not an LLM and it does not guess: each known setup is a checklist, and Whisper fires only when every box clears: structure, tape agreement, session context, and enough clock left to reach the target. Most sessions it fires a handful of times or not at all. Silence is a feature.

The five archetypes

👁 WATCHING, silence with reasoning

When nothing has fired, Whisper shows the closest setup per symbol with its checklist score ("3/4 met") and exactly which condition is still open. You always know why it is quiet, and which box the market still has to tick. Watching cards are not entries; they are the pre-flight board.

Signal lifecycle, read the stamp before you act

Every fired card carries a live status so nobody jumps into a stale signal:

• 🟢 FRESH, fired inside the last 10 minutes, spot still near the entry.
• 🔵 ACTIVE, still live; check spot against the signal entry before acting.
• 🚀 IN FLIGHT, already 40%+ of the way to target. The entry window has passed, do not chase.
• 🎯 TARGET HIT / ⛔ STOPPED, resolved. The card is a receipt now, not an entry.
• 🌙 SESSION CLOSED, the 0DTE session ended before either level traded.

📊 The Track Record, receipts not promises

The strip at the top of the Whisper tab is the honest tape: every fire is persisted the second it happens, the outcome (target vs stop on the underlying) is stamped at tick time, and a paper account mirrors every signal with real sandbox fills, entries at fire, exits when the level trades, flat by 15:55 ET. Wins, losses and session-closes are all counted; if a gap crosses both levels between ticks, it counts as a stop. The scorecard breaks it down per day and per archetype, so you can see not just whether Whisper wins, but which setups carry the edge.

Reading the confluence cards

The dashboard shows compact cards, symbol, direction, score, R:R, target, stop. Tap any card for the full detail modal.

The card color border = direction (green for LONG, red for SHORT). The score chip color tier: 7+ = high-conviction, 5-6 = mid, below 5 = doesn't show.

Index Sync, when SPX/SPY/QQQ agree (or don't)

Index Sync measures whether the three major index ETFs' dealer-positioning structures agree on bias. Five possible states:

The Tape Aligned state is a "tape sanity gate", it prevents the confusing "structure says Diverged but price is clearly rallying" reads we used to see.

Reading the Track Record

The Track Record section on the dashboard shows the live engines' actual performance, no backtested fantasy, no cherry-picked windows. Two panels:

Sample size matters. The engines have only been live since mid-June 2026, so most buckets are still accumulating. Numbers will firm up over the next 2-3 weeks of trading.

Mantis, the swing agent LIVE · graded daily · in testing

Mantis is a patient swing-trading agent for single names and the index ETFs, named for the predator that waits motionless and strikes only when the target is in range. It hunts multi-day option swings the way the desk does: dealer positioning first, everything else as confirmation. It is advisory. It surfaces the idea, the plan and the evidence; you pull the chart and decide. No card is ever meant to be taken blind.

Receipts so far: ETHA +141%, AAPL +70%, MSFT +63%, META 613 → 647, and this week QQQ +96% on the Sep 25 720 call after the card said "long as long as 700 holds" through the FOMC drop. Those are the hits. The misses are in the same journal, next to them, and the journal is what you should read.

How a card gets made

1. Universe. ~130 names: the discovered movers, every theme cluster (mega caps, memory, AI infrastructure, nuclear, crypto, small caps...) and the index constituents. Leveraged wrappers (TQQQ, SPXL, SOXL and their kin) are excluded: their chains are too thin to read and their thesis is the 1x thesis.
2. Technicals. Weekly and daily trend, RSI, ATR, relative volume, and one of five setup archetypes: trend pullback, breakout + retest, base breakout, MA reclaim, failed breakdown (and their short mirrors).
3. The Gravity gate. The dealer map across five expiries plus the ticker's own darkpool block walls and the week-over-week positioning shift. No dealer thesis, no card, ever. The heaviest structure below spot (with a block wall behind it) is a long; the mirror is a short.
4. Tilt. The share of dealer gamma sitting below versus above spot, weighted so a node at price this week outvotes a far, far-dated wall. A book split 45–55 is balanced and is withheld: no card, listed in the "Withheld as balanced" drawer with its structure read so you can still see the map. A card whose direction runs against the heavier side carries a GAMMA AHEAD flag.
5. Confirmation score, 0 to 100. Gravity 35, technicals 25, institutional flow 20, sector rotation 12, contract quality 8. Penalties after: earnings inside the hold window −15 (and a hard flag), thin contract −10, weekly-trend conflict −10, runner R:R under 1.2 −8, macro events −5. 55 is the floor. Below it the idea is journaled, never shown, except in the muted "below the floor" strip.
6. The contract. A 14–25 day expiry, 0.30–0.50 delta, liquidity-checked (open interest, spread). A setup with no tradeable contract is a chart, not a trade.

The plan on every card

• Entry is the price at the time of the read, with the archetype's trigger in words ("on strength through 709.92 while the 10-day low held").
• Stop is 1.2 ATR from entry (volatility-scaled, the dashboard's 43% → 61% lesson). Invalidation is a daily close through it. A wick through the stop is a warning the board flags as STOP TAGGED; the close decides, and if it closes through, the exit is that close, so a loss can exceed 1R. Gap risk is yours.
• T1 and T2 come from a ladder of real structure ahead of the trade: dealer nodes (magnets and gatekeepers, from the sign of their gamma), darkpool walls, and vanna nodes (VEX, where dealer delta hedging accelerates as implied vol moves with price). T1 sits where most cards actually travel (about 0.75 ATR), T2 is the runner (about 1.75 ATR), capped at what the horizon can plausibly deliver. Each target names its source and role.
• 1R in dollars is printed on every card: entry, stop, per-share risk, percent of entry. Every R on the card is measured from that entry and that stop, never from a floor below spot.
• The management plan: take 50% at T1 and move the stop to breakeven, 30% at T2, trail the last 20% by 1 ATR. That is what turns a high T1 hit rate into protected R instead of a coin flip.
• What the plan pays on the reference contract: the premium at T1, T2 and the stop for the named option, delta-only, so it understates on the way up. Same trade, in the units you actually trade.
• The contract is an example, not the trade. Mantis names one liquid 14 to 25 day, 0.30 to 0.50 delta contract so the plan can be graded in premium and so you can see what 1R looks like in option terms. The levels are the product. Take the strike, expiry or spread you are comfortable with: an expiry at least a week past the hold window survives a slow start, higher delta tracks the underlying and bleeds less, a vertical caps the loss. Size from the "at stop" line: contracts = (account × risk per trade) ÷ (premium × 100 × stop loss %). Risking 1% of a $25,000 account on a $9 contract whose stop is minus 73% is one contract, not five.
• The other side: the bear case for a long (or bull case for a short) in two sentences: the heavier structure on the far side and the close that proves the card wrong.
• Long-dated structure: the top nodes in the trade direction from the heaviest expiry beyond 60 days, with how many expected moves away they sit. A node beyond two expected moves is marked dormant: context, never a 7-day target.
• The structural read: Holding, Capping, Blocks, Vanna, Weight and a one-line Read, the way a desk describes a book. Roles come from the sign of each node in the end-of-day book; at crowded strikes the sign can be flipped, so read the role as a hypothesis and let price confirm it.

Reading the board

The Mantis tab is a desk blotter, not a card grid. The strip on top says what the book is: open ideas, net direction (a 100% long book is a warning, not a signal), index-equivalent exposure (SPY plus the heavyweights that move it are one bet), and data flags. Rows sit in four lanes by what they need from you: At the zone (price is in the entry band: the only rows that need a decision today), Armed, not triggered (plan set, price has not arrived), Fresh, unconfirmed (flagged after the close, wants one session of confirmation), and Gone without you (it ran past the entry; chasing here is how a good setup becomes a bad trade). Tap a row for the plan, the evidence, the contract, the structural read, the other side, and any warnings: ILLIQUID, TARGETS INVERTED, EARNINGS, DUPLICATE (same index in two wrappers), STOP TAGGED, GAMMA AHEAD.

A real card from this week

What happened: the market sold off through the Fed decision that afternoon, the 700 floor held on the close, and T1 printed the next morning. Read a card top-down in the order the agent decided: who, the plan, what proves it wrong, the evidence, the vehicle. Then open the chart. The card tells you where the structure is; the chart tells you whether price respects it today.

Cadence, lifecycle, and how it learns

Five scans a day (8:45 pre-market, 10:15 open drive, 1:00 midday, 3:05 power hour, 4:40 post-close), a monitor every 15 minutes in between, and the close-based stop settlement after the bell. A card lives LIVE → TRIGGERED → T1 ✓ → T2 / closed through stop / expired; the plan you were given is never rewritten by a later scan (a refresh may update the judgement, never the entry, stop or targets), and a card retires with a reason when the thesis flips or the setup relocates by more than an ATR.

Every candidate in every scan is journaled, including the ones that never became cards. Every retired card is graded on daily bars, and every named contract is graded minute by minute ("could you have taken profit?"). Two stop rules are graded side by side on every card (wick-through versus close-through) and the ATR stop is graded against a structural stop (the next node past the floor, minus 1%), so the rule that stays is the one the record supports. The Setup journal on the tab cuts the graded record by ticker, setup, volatility, direction, horizon and gamma tilt, and sizes each cut FULL / HALF / STARTER / AVOID once it clears six trades. Kairos, the inference engine, reads the same cards and outcomes and is learning a calibrated probability for each one.

Why "in testing": the scoring weights are a prior, not a finding, until the journal has enough trades to re-fit them (the first 100 is the bar). Until then the curated two or three strongest cards a day go to the Mantis channel in Discord with the full plan attached, and the tab itself opens to members as it earns it. Pull the chart. Size like the number could be wrong, because on some days it is.

Discord alerts, what fires & when

Every alert channel is gated for high conviction and freshness. Goal: ~5-10 actionable pings per RTH day total, never a spam stream.

The 10 House Rules

The map gives you edge. These rules keep you alive long enough to use it. Capital first; every rule below derives from that one.

1. Capital first. Edge second.

   Protect what you have before you try to grow it. No setup is worth blowing up the account that takes the next setup. Position size assumes you'll be wrong sometimes.

2. Trade the edges. Skip the middle.

   Fade reversals at WALLs, FLOORs, and KING magnets. The middle of the range has symmetric risk and no edge. If you're not at a structural level, you're guessing.

3. Asymmetric or pass.

   If the trade isn't at least 2:1 reward-to-risk from a clear structural level, don't take it. Win rate doesn't matter if you bleed on losses. R:R is the math that compounds.

4. Confluence beats conviction.

   A KING with a darkpool print at it, plus Index Sync agreement, plus price action confirming, that's a real setup. A KING alone is a hypothesis. Layer signals before you size up. (This is exactly what the Quick / Multi-Week confluence engines score.)

5. Don't fight reshuffles.

   When the map turns dirty (MAP_DIRTY transition fires), levels are unreliable. The right move is to pause. Re-engage when the structure settles (MAP_STABLE). Patience compounds.

6. Honor calendar hazards.

   Monthly OPEX, Liquidation Window (3:30 PM ET onward), earnings, FOMC, these distort positioning. Use lighter size or stand aside until clarity returns.

7. Long fear, short greed.

   Mean-revert at the extremes. Red candles into a FLOOR with positive GEX = buy. Green candles into a WALL with darkpool sellers = fade. The crowd is most wrong at the edges.

8. Size matches confidence.

   8/9 confluence + Index Sync aligned + Hawks confirming = standard size. 5/9 with one missing signal = half size. Diverged or shifting = no size. Never max-size on hope.

9. Don't let a green trade go red.

   If a trade gets to 1R in your favor, move stop to breakeven. If it gets to 2R, take partial. Markets give back winners that aren't protected. The best loss is the one you avoid by exiting.

10. Walk away after impact.

    After a big win or a big loss, your judgment is compromised. Either way, the next trade is the most dangerous one. Step away. Come back when you're flat-minded.

Next step: the Setup Library

The Playbook explains the system. The Setup Library shows you the five named dealer-positioning patterns to recognize visually on the chart, what they look like, when they show up, and the trade behavior for each.

Open the Setup Library →