gtrade.io

wallet · real · read from chain

$146.09

Bbhk…vxg2 · 1.6309 SOL · 6 tokens · 6 unsellable $576.63

EXECUTOR · watchingevery trade gets a reason
paper equity
$525.76
trades
5
wins
2
losses
3

Live flowscan → check → size → exec → log

checked 57881
  • $ALONHOLDcooldown 16m after 3 losses
  • $buttFILTEREDtop 10 hold 35%
  • $$MFRGFILTEREDdev holds 14%
  • $WEALTHSKIPbuys 50% of flow [momentum]
RUG CHECKFILTERED $butt
pass → size → enter
EXECUTORarmed
Filtered6113 skipped
Sizer—waiting
Loggerline #4317each with a reason
Wallet$146.091.6309 SOL · real

Trench feedcreated → filling → graduated

polling every 1s · 1.0 Hz
Parsed8,951+30 in the last minute
Graduated498pool opened this session
Last tick50 polls · 50s94 new · 67 filling · 61 grad
Judged2452 candidates
Graduate lane3 clean / 601 signalled · 2–90m after graduation
Attention36 hot · 22 called6 new to the desk · 25/25 tape wallets scored
Search funnel245judged17clean2signalled0DD pass0entered
rug checkmarket cap too small70rug ratio40liquidity below floor30top-10 concentration25bundler share19
signalthin flow6volume below floor5buy share265 swaps since launch, thin15m momentum1
  • $PUMP THIS4vtw…pumpnew4s
  • $keHpFy…pumpnew8s
  • $STARSHIPSKHZ…pumpnew12s
  • $TPOY44yq…pumpnew12s
  • $puterGM38…pumpnew13s
  • $ZPIX4ASR…pumpnew17s
  • $WINGBUDDY84AB…pumpnew16s
  • $JPEGZGGyd…1dp7new20s

Tradesopen book · price since entry

0 open5 closed

no open trades · the executor waits for a PASS

$piLLlonggraduatedev 60 unprovencalls weakhot #1closed

$43.52 · 8% of equity · lasted 39s

−0.6%−$0.27
entry
$9.019e-5
exit
$8.964e-5
stop
$8.117e-5 (-10%)
tp1
$0.000117
high water
+1.4%
marks
34 ·

reason: graduate: graduated 11m ago · 658 holders · buys 55% · 2.2× grad mcap · strong cluster: 1 smart buys in 30m · tape 7 wallets ×8.7 weighted (avg track 70/100 over 7/7) · $1,166 in · call-outs weak 38: 1 caller, 33% still held, 1 sold into the call · heat 82/100 · hot-search #1 (283 visits) · 22.1% eq by pool share · holders light ×0.75 · chief ×0.5

chief: Clean 86 DD, 11m graduate, smart-money cluster and #1 heat support the lane; but +92% 5m spike, fading volume, thin $23k liq and bundlers near cap argue half a ticket, not full. (62%)

exit: escape 70: 9 smart wallets fully closed in 15m

  • $NESTgraduate · trail 8% hit -37.9%−$11.37 · -38.6%
  • $CASSOsmart-money · time stop 90m, only 0.8%−$0.04 · -0.1%
  • $FUSEsmart-money · trail 15% hit 17.0%+$30.98 · +16.7%
  • $GWgraduate · escape 70: 2 smart wallets fully closed in 15m+$6.46 · +10.4%

$ tail -f night.log

4317 lines
  1. HOLD$ALON cooldown 16m after 3 losses
  2. FILTERED$butt top 10 hold 35%
  3. FILTERED$$MFRG dev holds 14%
  4. SKIP$WEALTH buys 50% of flow [momentum]
  5. SKIP$trauma 5m -2.5%, no momentum [momentum]
  6. FILTERED$HOLDMAS liq $3k < $5k
  7. FILTERED$stockloan mcap $7k too small
  8. FILTERED$BORDERLESS mcap $32k too small
  9. FILTERED$LOAD rug ratio 31%

Resultbalance curve · paper

one night · → now

balance · start $500.00 · max 10% per trade

$525.76

+$25.76 (+5.2%)

realized $525.76

1.05×

2W / 3L · worst −$11.37

wallet on-chain · 1.6309 SOL · 6 tokens · 6 unsellable $576.63 · read $146.09

AI deskchief decides · analyst vetoes · reviewer proposes

Chief· final call, pass or shrinkdeepseek-v4.1-flash

no ticket reached the chief yet · it decides only after rug check, due diligence and the analyst all pass

Analyst· veto onlygpt-5-mini

no candidate reached the analyst yet · it only sees tokens the rulebook already approved

Reviewer· morning note, proposals onlygpt-5.1-thinking

the note is written when a night ends (stop, halt or kill). Ask for one now to review the session so far.

Go-live deskPILOT gate · execution test, real but tiny

unlockable

paper record · all nights

14
trades
4
nights
50%
win rate
+22.4%
avg win
-12.7%
avg loss
+4.83%
expectancy
1.89
profit factor
3.6%
max dd
+$31.67
net

pilot gate · 3/3 cleared

  • ≥ 5 closed paper trades14
  • ≥ 1 paper night4
  • 1 position, ticket ≤ 20%pilot

env · complete

keys, wallet, PIN and budget cap present on the server · enter the PIN and run pre-flight to prove them against the API

arm · 8h window, real funds

wallet Bbhk…vxg2 · real1.6309 SOL · $146.09

Log80 of 80 shown

  1. HOLD$ALON reason: cooldown 16m after 3 lossesrisk
  2. FILTERED$butt reason: top 10 hold 35%rug check
  3. FILTERED$$MFRG reason: dev holds 14%rug check
  4. SKIP$WEALTH reason: buys 50% of flow [momentum]scanner
  5. SKIP$trauma reason: 5m -2.5%, no momentum [momentum]scanner
  6. FILTERED$HOLDMAS reason: liq $3k < $5krug check
  7. FILTERED$stockloan reason: mcap $7k too smallrug check
  8. FILTERED$BORDERLESS reason: mcap $32k too smallrug check
  9. FILTERED$LOAD reason: rug ratio 31%rug check
  10. SKIP$STM reason: vol $1k low [momentum]scanner
  11. SKIP$Diamond reason: vol $3k low [momentum]scanner
  12. FILTERED$puter reason: mcap $24k too smallrug check
  13. SKIP$duge reason: 82% below ATH, dead cat [momentum]scanner
  14. FILTERED$puter reason: dev holds 30%rug check
  15. FILTERED$porkchop reason: mcap $7k too smallrug check
  16. FILTERED$OTC reason: bundler 41%rug check
  17. FILTERED$computer reason: liq $3k < $5krug check
  18. FILTERED$putercat reason: graduated 1m ago, migration dumprug check
  19. HOLD$ALON reason: cooldown 17m after 3 lossesrisk
  20. FILTERED$PUTER reason: LP lock unverified on meteora_damm_v2rug check
  21. SKIP$trauma reason: 5m 0.8%, no momentum [momentum]scanner
  22. FILTERED$HOLDMAS reason: bundler 35%rug check
  23. SKIP$Rocco reason: buys 50% of flow [momentum]scanner
  24. FILTERED$puter reason: liq $3k < $5krug check
  25. FILTERED$SGAT reason: mcap $4k too smallrug check
  26. SKIP$PI reason: 9 swaps, thin flow [momentum]scanner
  27. SKIP$WHITE reason: 29 swaps, thin flow [momentum]scanner
  28. FILTERED$puter reason: mcap $3k too smallrug check
  29. FILTERED$COIN reason: bundler 31%rug check
  30. FILTERED$OTC reason: mcap $5k too smallrug check
  31. FILTERED$TA500 reason: mcap $11k too smallrug check
  32. FILTERED$computer reason: mcap $11k too smallrug check
  33. HOLD$ALON reason: cooldown 18m after 3 lossesrisk
  34. FILTERED$PEPE69 reason: liq $79 < $5krug check
  35. FILTERED$Ram-Altman reason: LP lock unverified on meteora_damm_v2rug check
  36. FILTERED$pruple reason: top 10 hold 63%rug check
  37. FILTERED$HOLDMAS reason: liq $3k < $5krug check
  38. FILTERED$11/13 reason: bundler 35%rug check
  39. FILTERED$OREN reason: age 222h, stalerug check
  40. SKIP$Rocco reason: 25 swaps, thin flow [momentum]scanner
  41. SKIP$Diamond reason: 38 swaps, thin flow [momentum]scanner
  42. FILTERED$モグラ reason: bundler 32%rug check
  43. FILTERED$COIN reason: mcap $35k too smallrug check
  44. FILTERED$plump reason: liq $4k < $5krug check
  45. FILTERED$ADDOG reason: dev holds 30%rug check
  46. FILTERED$Kirks reason: mcap $24k too smallrug check
  47. HOLD$ALON reason: cooldown 19m after 3 lossesrisk
  48. FILTERED$cp reason: mcap $212 too smallrug check
  49. FILTERED$puter reason: LP lock unverified on meteora_damm_v2rug check
  50. FILTERED$PUTER reason: LP lock unverified on meteora_damm_v2rug check
  51. FILTERED$Trump reason: top 10 hold 88%rug check
  52. FILTERED$Ram-Altman reason: dev holds 20%rug check
  53. SKIP$Rocco reason: vol $7k low [momentum]scanner
  54. FILTERED$computer reason: liq $3k < $5krug check
  55. FILTERED$BESEECHER reason: mcap $7k too smallrug check
  56. FILTERED$Ram-Altman reason: mcap $6k too smallrug check
  57. SKIP$Kirks reason: mcap $28k = 0.69× graduation, bleeding below the curve [graduate]scanner
  58. HOLD$Rocco reason: cooldown 20m after 3 lossesrisk
  59. FILTERED$WSB reason: mcap $14k too smallrug check
  60. FILTERED$PEPE69 reason: liq $0 < $5krug check
  61. FILTERED$PEPE69 reason: dev holds 44%rug check
  62. FILTERED$MAG7 reason: top 10 hold 57%rug check
  63. FILTERED$ALON reason: LP lock unverified on meteora_damm_v2rug check
  64. FILTERED$PUTER reason: dev holds 19%rug check
  65. FILTERED$HERO reason: LP lock unverified on meteora_damm_v2rug check
  66. FILTERED$PUTER reason: dev holds 19%rug check
  67. FILTERED$PUTER reason: LP lock unverified on meteora_damm_v2rug check
  68. FILTERED$NIKE reason: liq $2k < $5krug check
  69. FILTERED$ALON reason: liq $3k < $5krug check
  70. FILTERED$xat reason: mcap $4k too smallrug check
  71. FILTERED$TA500 reason: top 10 hold 50%rug check
  72. FILTERED$porkchop reason: bundler 35%rug check
  73. FILTERED$terminal reason: mcap $3k too smallrug check
  74. FILTERED$Kirks reason: mcap $17k too smallrug check
  75. HOLD$Diamond reason: cooldown 21m after 3 lossesrisk
  76. FILTERED$butt reason: liq $3k < $5krug check
  77. FILTERED$miggles reason: liq $0 < $5krug check
  78. FILTERED$ALON reason: dev holds 11%rug check
  79. FILTERED$su reason: liq $7 < $5krug check
  80. FILTERED$HERO reason: LP lock unverified on meteora_damm_v2rug check

Rulebookhard limits, not knobs

Sized for a $100 book. Three positions at once, each 33% of equity, because the fixed fee is per swap: five $20 tickets pay 6% of themselves in round-trip cost, three $33 tickets pay 4.4%, and the aggregate risk is the same 10% either way. A $100 book therefore opens $33 tickets, and the profile needs $40 before it opens anything at all. Winners run to +30% against a 10% stop, so a 60% win rate is profitable instead of needing 70%. The night survives three stop-outs (-10%) and halts at -15% from peak. Fraud filters, the dev-score floor, the direction gate and the circuit breaker are unchanged. Risk-off still blocks entries, the escape monitor still sells first, and live stays behind the paper gate.

SCANNER · parse everything, judge everything

  • trench feed polls GMGN launchpads for ~50s of every minute (1 Hz on Plus): every new creation, curve fill and graduation is parsed and stored
  • universe per tick = 5m + 1h trending boards (100 rows each) + every token that graduated or is about to
  • two lanes: board rows are judged by the board rulebook, graduates by the graduate lane — a board row that just graduated goes to the graduate lane
  • rug-check and signal verdicts are re-judged every tick; only entered / vetoed / no-go stick for the session

RISK · survive first

  • 3.3% of equity at risk per trade, position ≤ 33%
  • max 3 open, ≤ 100% deployed, ≥ 2m between entries
  • position ≤ 0.5% of pool · est. slip ≤ 2%
  • round-trip cost ≤ 8% of the ticket — the fixed fee is per swap, so a ticket below $14 is refused however good the signal
  • no entries below a $45 balance — under that every ticket the profile allows is too small to clear the fee cap
  • night ends at -10% or -15% from peak
  • 3 losses in a row → 60m cooldown

RUG CHECK · hard filters

  • mint + freeze renounced, no honeypot, tax ≤ 5%
  • rug ratio ≤ 30%, top 10 ≤ 35%
  • liq ≥ $5.0k (derived: the $14 fee floor at the 0.5% pool cap, not a constant), mcap $40k–3M, mcap/liq ≤ 20x
  • holders ≥ 120, age 5m–72h, LP burned or locked
  • bundler ≤ 30%, insider ≤ 20%, dev ≤ 10%, snipers ≤ 20%

GRADUATE LANE · launchpad graduates, same desk

  • Fresh graduate: Bonding curve completed minutes ago, pool holding above graduation after the migration dump while buyers keep arriving. Enter after the dump, exit mechanically before the crowd leaves.
  • rug profile: liq ≥ $5.0k, mcap $25k–2M, holders ≥ 80, 2–90m after the pool opened
  • authorities, tax, top 10, bundlers, insiders, dev, snipers, mcap/liq and LP burned/locked: identical to RUG CHECK — a migrated pool with an unverified lock is skipped
  • signal: ≥ 200 swaps, vol ≥ $12k, buys 48–92% of flow, mcap ≥ 0.8× graduation level (~420 SOL)
  • then the same desk as every candidate: ranked by score, DUE DILIGENCE, SIZER, ANALYST, CHIEF call, EXECUTOR, ESCAPE monitor

STRATEGIES · ranked, best one trades

  • Momentum continuation: Fresh buy pressure on a token that is rising but not vertical. Ride the next leg, exit mechanically.
  • Smart-money follow: Wallets GMGN tags as smart/renowned are accumulating while price is still calm. Follow the informed flow, not the crowd.
  • Pullback reclaim: Strong 1h trend, shallow 5m pullback that buyers are absorbing. Enter the dip inside a trend, never a falling knife.
  • shared floor: buys ≥ 54%, ≥ 40 swaps, vol ≥ $12k or ≥ $600/min (lifetime volume over a 20m window, so a young hot token is not judged as a stale one), 5m +1–45%, 1h ≤ +200%, ≤ 70% below ATH

DUE DILIGENCE · GMGN skills, deterministic

  • rank = 60% strategy + 40% trend-momentum model + cluster bonus (≥3 smart buys in 30m) + call-out bonus ≤ 10 + heat bonus ≤ 10
  • cluster wallets weighted by their 7d wallet-score track record (0.4×–1.6×, unscored 1×; 6 new wallets scored per tick, cached 6h); smart money net selling in 30m → no entry
  • hot-search list (1h) joins the universe; heat rank = percentiles over the tick's universe (volume · holder growth · quality · smart money · ATH position · turnover+visits · hot-search)
  • call-outs (last 120m): callers · KOL · still holding · caller 30d hit-rate (3 callers checked); callers who sold into their own call → ticket ×0.75, bonus withdrawn
  • top 4 only: token security re-read (honeypot, mint, freeze, tax, top10, insiders)
  • dev score ≥ 15/100: launch survival rate, stuck-inner factory penalty, logo reskins, exited dev
  • dev CONDUCT ≥ 50/100 from his own trades in his own coins (40 rows): pull multiple × first-sell speed, dump gate (≥1.5× within 30 s), no record reads 60
  • 1m chart over 60m: pattern score ≥ 30, no breakdown / bleed / distribution, < 5 red in a row
  • composite DD ≥ 50/100 (contract-dd weights: security 45 · holders 35 · chart 20, worst row rules), lane-aware for fresh graduates
  • bundler-sell signal in the last 30m → no entry
  • direction gate at the fill: 5m ≤ -10% blocks the market buy, ≤ -5% halves it; graduates read the 1m bar
  • slippage per order: 2% + tax + 1.5× price impact + ½|Δ5m|, capped at 12%

ESCAPE · leave before the exit is gone

  • every fill re-read each tick: honeypot raised, mint re-acquired, top10 +15pp, liquidity -30% → severity ≥ 70 sells first
  • top-100 holders re-read every 5m on open positions: rat traders, sock puppets and a whale > 10% raise severity
  • ≥ 2 smart wallets fully closing inside 15m → exit; smart money selling into KOL buys → leash
  • volume collapse (last 20 bars < 0.2× the 20 before) → leash: trail 8% from high-water now, 30m clock

REGIME · do not fight the tape

  • breadth = share of trending tokens green on 1h
  • risk-on ≥ 55% green: full size
  • neutral: size × 0.8
  • risk-off < 40% green or median 1h < -10%: no new entries

EXIT · mechanical

  • stop -10% from fill, set before entry
  • tp1 +30%: sell half, stop to breakeven
  • then trail 18% below high-water
  • time stop: 45m without +8% → out
  • live: every buy carries a GMGN strategy — tp1 half at +30% and a -13% stop; the strategy id is verified on the next tick, a missing or stopped strategy is re-placed or flagged bot-only
  • live: once GMGN fills tp1 the strategy is replaced by a break-even stop (+1% over open), so a dead engine still cannot give back the trade

AI + LIVE GATE · powerless by design

  • ANALYST (LLM) sees only rulebook passes; its veto is final at ≥ 90% confidence, below that it is a DISSENT and CHIEF may still enter at ≤ ×0.5
  • CHIEF (DeepSeek) makes the final entry call: pass, or enter at ≤ the proposed size
  • LIVE_GATE=pilot: live execution test after 5 paper round trips · 1 position, ticket ≤ 40% of the selected profile, budget capped by LIVE_MAX_BUDGET_USD
  • REVIEWER (LLM) writes the morning note; proposals, never edits
  • live unlocks after ≥ 30 paper trades over ≥ 5 nights
  • expectancy ≥ +1%, PF ≥ 1.3, max DD ≤ 10%
  • PIN + budget cap + 8h arming window + kill switch
  • swap circuit breaker: the same business error 2× in a row pauses live entries 30m before GMGN blocks the key; transient errors do not count
what scaling changes against guarded
  • board: liq ≥ $12k (was $30k), mcap ≥ $40k (was $150k), holders ≥ 120 (was 300), age ≥ 5m (was 15m)
  • graduates: liq ≥ $6k (was $10k), holders ≥ 80 (was 150), window 2–90m (was 3–45m)
  • board signal: buys ≥ 54% (was 56%), ≥ 40 swaps (was 60), vol ≥ $12k (was $25k), 5m +1–45%, 1h ≤ +200%
  • graduate signal: ≥ 200 swaps (was 300), vol ≥ $12k (was $20k), buys 48–92%, hold ≥ 0.8× graduation (was 0.9×)
  • due diligence chart score ≥ 30 (was 45), composite DD ≥ 50/100 (was 70), dev score ≥ 15/100 (was 15), dev CONDUCT ≥ 50 (was 50); security re-read, holder health and bundler-sell gate unchanged
  • paid checks on the top 4 candidates per tick (was 3)
  • executor slippage cap 12% (was 10%); direction gate (5m ≤ -10% blocks, -5% halves) unchanged
  • analyst veto is final only at ≥ 90% confidence; a weaker veto is a dissent and the chief may enter at ≤ ×0.5
  • neutral regime size ×0.8 (was ×0.6); risk-off still blocks entries
  • max 3 open (was 2), 2m between entries (was 10m); risk per trade, caps, stops, breakers unchanged