Settings

dry_run records copy decisions without touching Maestro; live drives the real Operator Finder Telegram account.

Copy slots the engine may hold open at once — each slot is a tracked burner in Maestro.

Burners tracked per play for an operator with no per-operator override.

A slot whose burner has gone this long without activity is closed as stale.

Per-buy ETH ceiling in live mode — the hard cap under every Maestro buy this engine causes.

The exact chain-button text Maestro shows for Robinhood (e.g. HOOD). Empty keeps live mode refusing to act.

Master switch. Off = the nightly cycle previews (logs what it would do) and moves no wallets.

Floor-only Sim->Live promotion (#2472). With auto_execute_enabled, the nightly pass promotes any sim wallet whose score AT DECISION TIME clears live_min_score — and asks nothing else about its record. The evidence is the corrected #2452 report: >=70 is the only floor positive at $12 and it held +$17.5/trade through a book-wide -$7.7k stretch. Still gated on capability (can_execute_sell_all, never overridable), sim tenure, and the per-chain live cap. The MANUAL gate's trades/days/win-rate/PF/30d rules are untouched and do not apply in this mode. Ships OFF.

Daily funnel digest + urgent live-demote ping (#672).

Simulation-lane slots. Paper only — the cost of raising it is API/compute budget and list length, not money.

Live-lane slots PER CHAIN (#2450), enforced by the promotion endpoint. Keep equal to Maestro's per-chain copy-trade slot count (5 free, 12 premium): below idles earned slots, above lets a promotion succeed here and fail to sync. Per chain because Maestro's slots are per chain and `sync_wallets` already pushes per chain — a global cap refused a sixth live wallet even onto a chain with all five slots empty.

Per-chain sim slot reservation (#1327): each reserved chain keeps this many slots only its own candidates can fill.

Comma-separated chains holding a sim_chain_floor reservation. Empty reservations cost nothing.

Throughput bound on the candidate pool (#2227) — NOT a merit valve (#2198/D3): eviction is labelled pool_cap, makes room by rank, and 0 disables.

When the pool is full, discovery stands down instead of competing for the Gecko budget the tracker needs (#1831). 0 = keep scanning regardless.

Candidate→Sim: settled trades required before the gate judges.

Candidate→Sim: days of tracking required.

Candidate→Sim: minimum win rate (%). The measured cliff sits at 40% — wallets above it hold 96.8% of book PnL.

Candidate→Sim: maximum rug rate (%).

Candidate→Sim: minimum lifetime net PnL (USD).

Candidate→Sim bot detector (#2271): above this settled-trade RATE the wallet runs a cadence our copy path cannot mirror (FERRET: 186/day, 18% win, −$2,891 simmed).

Sim→Live: settled sim trades required.

Sim→Live: days in the sim lane required (#2170: the sim-lane clock, not lifetime tenure).

Sim→Live recency floor (#1259): the last 30 days must have earned this much (USD) — lifetime-positive is not enough.

Sim→Live quality floor (#1305), passable EITHER by win rate OR by modern profit factor (#1933) — a 2x-fib ladder is many small losers and few large winners by design.

Sim→Live: the OTHER way past the quality floor (#1933). 2.0 = wins two dollars per dollar lost.

Sim→Live: modern-era sample required behind the profit-factor branch.

Sim→Live: maximum rug rate (%).

The grade line of the sim→live gate — BINDING since #2386: live requires evidence AND grade, ungraded fails. Stored value governs; the operator holds it at 60 (B−) as of 2026-08-16.

Read the three absolute-dollar thresholds at the reference notional (#2071), so a $5-stake live wallet is not judged on a $100 ruler. Ships off; flipping it is the operator's call.

The notional the dollar thresholds were written against ($100, from #740).

Composite weight: 30d net PnL vs its benchmark.

Composite weight: win rate on the floor/target ramp.

Composite weight: average PnL per settled trade.

RETIRED (#2034) — edge left the composite, so this weight is never applied whatever it is set to. Kept so stored configs still parse.

Composite weight: profit factor (#1902). Ships 0 = inert; the breakdown shows the figure either way.

Composite weight: ROI on staked capital (#1902). Ships 0 = inert.

Composite weight: sample size. Dropped automatically when shrinkage is on — scoring confidence twice is the bug.

Composite weight: max drawdown. Enters the score only when drawdown_scoring_enabled.

Composite weight: recency of activity.

30d net PnL (USD) that earns 100.

Average per-trade PnL (USD) that earns 100.

Edge PnL (USD) that earns 100.

Settled trades that earn the sample component 100.

Win rate (%) scoring 0 — the ramp starts here, placed below the measured 40% profitability cliff (#1589).

Win rate (%) scoring 100 — the top of the ramp.

Profit factor that earns 100. Neutral is 1.0 by arithmetic, which immunizes this benchmark against drift.

ROI (%) that earns 100. Normalized by stake, so sim and live grade on one scale.

Which era the performance components read: the whole record, or modern-entry rows only.

How a component scales to its benchmark. linear saturates at the benchmark; log keeps earning per 10x above it (losses stay linear). Reorders promotions — decide against measured impact (tools/score_curve_impact.py).

Under score_curve=log: the multiple of the benchmark that reaches 100.

Thin-record handling (#1579): pull small samples toward score_prior instead of scoring absence as badness; drops the sample component to avoid double-counting confidence.

The score a record with no evidence reads (#1592 d3: 50 = C on the rebased bands). Coupled to the grade scale — neither moves alone.

Settled trades at which the record carries half the weight against the prior.

BEST-hero eligibility score. Display rank only, never a gate. 60 sits a full band above the 50 starting line.

Let max drawdown enter the composite. Refused while coverage is below min_drawdown_coverage_pct (#2077) — scoring a component with no input makes the denominator wallet-dependent.

Coverage (% of wallets with computed drawdown) required before drawdown scoring may be enabled. 0 disables the gate.

RETIRED (#2034) — INERT IN BOTH POSITIONS. Edge no longer enters the composite for any wallet: it says which exit strategy suits a wallet, not how good the wallet is. Still measured and shown, with its sign and coverage, on every live-shortlist surface.

Edge is rankable only when at least this % of the book produced a mirror leg (#1520).

Edge is rankable only on at least this many scored positions — 100% coverage of 4 rows is not evidence.

Edge is unrankable when one position carries more than this % of the measurement (#1899). 0 disables.

Window (days) over which sustained underperformance is judged.

Settled trades required before demotion floors judge — a sample gate, not a floor.

Demotion floor: win rate (%) below this over the window (with sample) demotes. Deliberately far below the promotion floor.

Demotion floor: rug rate (%) above this demotes.

Demotion floor: window net PnL (USD) below this demotes. Calibrated on the $100 sim notional — see gate_notional_normalize.

Grade floor for demotion (#2195 companion). Shadow unless demote_min_score_enabled.

Enforce demote_min_score. Off = computed, logged, shown, and nothing moves (the shadow pattern).

Instant demotion: daily loss (%) that trips the panic path, no demotion window needed.

Instant demotion: consecutive rugs that trip the panic path.

Days quiet before a wallet reads as slowing down.

Days quiet before a PROVEN wallet is demoted as dead.

Shorter leash (#1584): days a wallet that has NEVER traded may hold a sim slot.

Flat idle decay (#2196/D2): ranking-score points bleed per whole day the LEAD shows no observed activity. Read-time only; replaces the recency term's idle contribution, never stacks.

Points off the ranking score per idle day.

Allow displacement swaps at all.

Competitive displacement (#1584): a queued candidate takes the weakest incumbent's slot only when it outscores by this margin — the anti-thrash guard.

Propose instead of execute (#1667). Off since #2246: nothing could approve a proposal, so the lane could not turn over.

Candidate eviction (#795): no new buys in this many days evicts.

Leeway for proven (ever-live) earners before inactivity evicts.

Name fresh sim promotions with a pronounceable handle (#1587). Never overwrites a real name (#1842); runs after the commit and cannot block a promotion.

Seconds the naming CLI call may take (#2242: 15 killed every call mid-answer; 60 is the measured-safe value).

Model for the naming call. Empty = the CLI's own default.

Scan Interval (seconds) — Time between automated scan cycles. Default: 3600 (1 hour)" [wallet_scanner.html:947-949]

Auto-Start Scanner" — card note: "Configure services to start automatically when the application launches. Delays allow time for testing/development." (ws) / "Configure services to start automaticall

Delay [30 min]" — no separate hint beyond the slider end labels "Now" / "60 min" and the Auto-Start card note

Candidate Re-checks per Cycle — All candidates rotate through the daily re-check, stalest first, capped at this many per cycle (#1058). 0 = no cap." [wallet_scanner.html:1008-1010]

Enabled Chains — Connectivity + trading + scanner universe (#1242). Applies after a service restart. Stored in settings; secrets.json is the fallback when unset. Solana copy-trade rides its own stack

Scan Chains — Which chains the scanner spends API budget on (#1043/#1242/#1187). Unchecked chains are paused everywhere — discovery sweep, candidate re-checks, and queued/manual wallets (skipped with

Use DeBank scraping — Operator kill switch (#1252). Off disables DeBank scraping entirely; discovery falls back to the chains it can index without it." [copy_trade.html:2902-2905]

Max DeBank Pages — Max pages to scrape per wallet (~20 txns/page). Prevents excessive wait times on wallets with 1000s of transactions" [wallet_scanner.html:1135-1137]

Top N Tokens to Scan — Max trending tokens to analyze per chain" [wallet_scanner.html:963-965]

Top N Traders per Token — Max traders to check from Top Traders list" [wallet_scanner.html:968-970]

Min 24H Gain (%) — Only scan tokens with this % gain in 24h" [wallet_scanner.html:1031-1033]

Volume Threshold ($) — Min daily volume required to include token in scoring. Higher = stricter. Default $500 works for alpha snipers on fresh launches." [wallet_scanner.html:1120-1122]

Matched Tokens per Sweep (per chain) — Per-sweep budget of filter-passing tokens to trader-scrape (#1025). 0 = unlimited." [wallet_scanner.html:1003-1005]

Rank discovery by gainers — Order the scan by top gainers rather than the legacy ordering. Config-gated so it can be reverted without a deploy." [copy_trade.html:2886-2889]

Min PNL Multiplier (x) — PNL must be this multiple of buy-in (2x = doubled)" [wallet_scanner.html:1057-1059]

High Grade Threshold — Score at/above this fires the Discord discovery alert and the \"high grade\" log. It does NOT gate saving: every wallet scoring ≥60 (or thin-strong) is kept as a Candidate; bel

Max Token Age (hours) — Only scan tokens younger than this" [wallet_scanner.html:1036-1038]

Exclude Suspicious Data — Exclude tokens flagged as suspicious (capped at 10,000x) from PNL and scoring calculations" [wallet_scanner.html:1141-1145]

Skip Hyperactive Solana Wallets — Pre-fetch volume-band skip of bot-grade wallets — saves Helius credits (#1025)" [wallet_scanner.html:1014-1017]

Auto-Refresh Failed Tokens — Retry tokens that failed (not indexed, no data)" (options verbatim: "Enabled (retry once/day, max 7 days)" / "Disabled") [wallet_scanner.html:1125-1130]

Max Buy-in ($) — Only traders who bought with less than this (filters whales)" [wallet_scanner.html:1052-1054]

Detection Mode — Count per token or total" (options verbatim: "Per Token" / "Per Wallet") [wallet_scanner.html:1100-1105]

Buy Amount ($) — Buys under this are \"small\"" (section note: "Filters out bots and snipers that make many tiny trades. These are hard to copy profitably.") [wallet_scanner.html:1095-1097]

Buy Count Threshold — Flag if more small buys than this" [wallet_scanner.html:1090-1092]

Fibonacci Sell % — 50% sold at 2x, then this % at 3x, 5x, 8x, 13x, 21x..." (card note: "Configure theoretical PNL strategies. All entries normalized to $100 for comparison.") [wallet_scanner.html:117

Trailing Stop % — 50% sold at 2x, remaining uses this trailing stop" (card note: "Configure theoretical PNL strategies. All entries normalized to $100 for comparison.") [wallet_scanner.html:1173-1175

Default Display Method — Which method to highlight in wallet cards" (options verbatim: "Trailing Stop (rides runners)" / "Fibonacci Laddering (profit-taking)" / "Show Both (comparison)") [wallet_scan

Execution Enabled — Master switch for live on-chain sells" (card note: "Execute sells on-chain via DEX routers when Fibonacci levels trigger. Wallet private key must be configured in secrets.json.")

Dry Run Mode — Log sells without submitting transactions" [copy_trade.html:1853-1854]

Default Max Buy per Wallet (USD) — Single source of truth for live buy size — converted to native at the live price and pushed to Maestro per wallet. On Solana this is the TOTAL debit per buy: the Ma

Slippage (%) — Tolerance for swap price impact" [copy_trade.html:1862-1865]

Forced-Exit Slippage (%) — Looser min-out for forced exits (stops, liquidity, time rules) so an emergency sell is not blocked by its own protection (#925)" [copy_trade.html:1872-1882]

Blockchain Poll Interval — How often to scan the blockchain for new wallet buys" (options verbatim: "3s — Fast (~114k credits/hr)" / "5s — Normal (~68k credits/hr)" / "10s — Efficient (~34k credits/h

Live Position Poll Interval (seconds) — How often LIVE positions are repriced (#2252). Simulation keeps the 30s loop; only this lane pays. Costs one price call per unique live token per tick — at 2 l

Stale Poll Interval (seconds) — Poll cadence for positions already flagged stale — a frozen price loses nothing at a slow cadence" [copy_trade.html:2357-2360]

Near-Trigger Poll Interval (seconds) — Cadence for a position CLOSE to a decision — the trailing stop, the hard stop, or the next fib rung (#2252). The tier only ever tightens: nothing is polled slow

Auto-Start Copy Trade" — card note: "Configure services to start automatically when the application launches. Delays allow time for testing/development." (ws) / "Configure services to start automatic

Delay [30 min]" — no separate hint beyond the slider end labels "Now" / "60 min" and the card note quoted for copytrade_auto_start

Resume the scan queue on boot — On (default, #1225) a queue interrupted by a restart is picked up where it stopped. Off restores the #840 policy of starting clean — a queue that can never be resumed

Outage catch-up — Operator kill switch (#2057). On, a stale watermark RECORDS a gap, the tracker pass puts gap wallets first, and wallets outside the candidate universe (armed leads) get EXTRA catch-

Minimum Liquidity (USD) — Exit if pool liquidity falls below this absolute value" [copy_trade.html:2471-2474]

Pre-buy liquidity floor — Refuse a buy when the entry liquidity is KNOWN and below Minimum Liquidity (USD) — the same one number the Liquidity Protection card owns, so the pre-buy floor and the post-

V4 pre-buy sellability probe — Runs only where the V2 on-chain sim cannot (a Uniswap V4 token has no V2 pair), so the blocking gate is otherwise blind there. Three states rather than a boolean on pur

Max token age — skip buy / sell held bag — Don't enter tokens older than this at buy time; sell a held token once it ages past it." (card note: "Force an exit (and skip stale entries) on the clock, i

Max token age — skip buy / sell held bag — Don't enter tokens older than this at buy time; sell a held token once it ages past it." (shares its panel's hint with ct_max_token_age_enabled) [copy_trade

Max Tokens Per Wallet — Max open positions per wallet (0 = unlimited). Skips new buys when cap is reached." [copy_trade.html:2156-2160]

Model token taxes in SIM — Apply the token's measured buy/sell tax to simulated fills. Off makes SIM optimistic against the same token LIVE would trade, which is the one direction a simulator must no

SIM entry impact, base V4 (%) — added to a SIM entry booked at the pair quote on a Uniswap V4 pool. The median lead-fill-over-quote gap printed by tools/measure_entry_impact.py (#2538).

SIM entry impact, base V3 (%) — added to a SIM entry booked at the pair quote on a V3 pool. LOW CONFIDENCE: n=11 on 2026-09-15, IQR +8.6% to +93.1%. The median printed by tools/measure_entry_impact.py (#2538).

SIM entry impact, Solana (%) — added to a SIM entry booked at the pair quote on Solana. Measured 0% (the quote is the fill) (#2538).

Entry impact measured on (YYYY-MM-DD) — the date of the sample behind the impact numbers above. Past the max age they are not modeled (#2538).

Entry impact max age (days) — a measurement older than this is not applied; the row stamps no modeled impact until the tool is re-run (#2538).

Sell Strategy — Familiar through 13x, then tapers for moonbag" (default-selected option's dynamic desc; options verbatim: "90/10 Fib Taper (recommended)" / "Front-Loaded Decay" / "Fat Middle" / "Cust

% of Remaining — Moonbag never depletes" / "% of Initial — Fixed chunks, can fully close" (button pair calling setFibSellMode()) [copy_trade.html:2254-2266]

Sell per Level (%) — % of initial position to sell at each level after 2x (position can fully close)" (hint dynamic, id=fibSellPctDesc) [copy_trade.html:2277-2281]

Free Roll at 2x (%) — % of initial position to sell at 2x" (hint text is dynamic, id=fibFreeRollDesc) [copy_trade.html:2270-2274]

Trailing Stop (%) — Sell remaining if price drops this % from peak (after hitting threshold)" [copy_trade.html:2312-2315]

Hard Stop-Loss — Close if down X% and never hit threshold" [copy_trade.html:2321-2322]

Hard Stop-Loss (%) — Close position if down this % and never reached threshold" [copy_trade.html:2330-2333]

Profit Threshold (multiple) — Must reach this multiple before trailing/breakeven stops activate" [copy_trade.html:2306-2309]

Max hold duration — sell immediately — Close any position held longer than this, regardless of price." [copy_trade.html:2417-2426]

Max hold duration — sell immediately — Close any position held longer than this, regardless of price." (shares its panel's hint with ct_max_hold_enabled) [copy_trade.html:2422-2426]

Near-Trigger Band — enter (%) — A position within this distance of its nearest trigger joins the fast lane." [copy_trade.html:2375-2378]

Near-Trigger Band — leave (%) — And leaves it only past this one. A BAND, not a line (#2308's shape): on a single threshold a position sitting on it changes cadence every tick. Set below the enter va

Liquidity Exit — Exit when pool liquidity drops below thresholds" (card note: "Auto-exit positions when pool liquidity drops (anti-rug). Uses on-chain data from swap events at zero extra RPC cost.")

Liquidity Drop (%) — Exit if pool liquidity drops this % from entry (e.g., 80% = liquidity went from $10K to $2K)" [copy_trade.html:2477-2480]

Applies to chains — Comma-separated (e.g. solana,base) or all." [copy_trade.html:2441-2445]

Applies to modes — Live also caps Maestro-executed positions — start with Sim." (options verbatim: "Simulation only" / "Live only" / "Simulation + Live") [copy_trade.html:2431-2438]

Stale Price Warning (hours) — Flag a position stale once its price has been frozen this long (#848); Dead Token Detection then classifies it at its own threshold" [copy_trade.html:2351-2354]

Dead Token Detection (hours) — Mark as rugged if price unchanged for this many hours" [copy_trade.html:2336-2339]

Orphan timeout (minutes) — How long a LIVE position waits for a Maestro fill before it is orphan-closed. Too short and a slow-but-real fill is written off; too long and the book carries a position no

Post-Close Cooldown (hours) — Continue recording candles after a position closes to see what happened next. 0 = disabled." (card note: "Records 1-minute OHLC candles during active positions for trade

Bag ledger — Whose tokens a sell is allowed to spend when two positions share one wallet bag. Ships off; promote it once shadow lines from real live sells have been compared. Anything unrecognised re

$100 SIM twin for LIVE buys — Book a paper $100 simulation sibling next to every LIVE position, at the same detected entry and through the same SIM friction (tax modelling above), so the sim lane's c

Treasury manager ARMED (#2532) — OFF (the default) runs the manager DRY: it computes the transfer it would make and writes it to the treasury ledger with dry_run=1, and signs nothing. ON lets it sign real SOL transfers between the treasury and the DBot-facing trading wallet — and only between those two addresses, which are derived from the keys in secrets.json rather than configured anywhere. Read the dry-run ledger rows before arming

Treasury: float target (USD) — How much the DBot-facing wallet is allowed to hold. DBot escrows imported keys with no export and has taken a chain down indefinitely, so this is the maximum we are willing to lose to the vendor. A dollar figure, converted to SOL at the modeled price each tick; a tick with no price SKIPS rather than sizing a transfer off a guess

Treasury: sweep band (%) — Sweep to the treasury only once free SOL is this far ABOVE the float target. A band rather than a line (#2532): on a single threshold a wallet sitting near target transfers on every tick forever, paying a fee each time. The transfer restores the balance to exactly the target

Treasury: top-up band (%) — Top the trading wallet up only once free SOL is this far BELOW the float target. Same hysteresis argument as the sweep band; the transfer restores the balance to exactly the target

Treasury: gas reserve (SOL) — Left untouched in BOTH wallets. Covers transaction fees and rent-exemption; a wallet that cannot pay a fee cannot sell. A top-up the treasury cannot fund without breaching this is REFUSED, never part-filled

Treasury: tick cadence (minutes) — How often the float is evaluated. Bookkeeping, not latency-sensitive: the trading wallet drifting a few minutes past its band costs nothing, and a fast cadence just spends RPC reads. Floored at one minute

DBot lane ENABLED (#2512) — Master switch for writing copy tasks to DBot. OFF = no task is created or edited, whatever a wallet's executor says. A wallet reaches DBot only when this is on AND that wallet's executor column is 'dbot'; nothing in this repo sets the latter — the operator does, one wallet at a time

DBot: Solana execution wallet address — The dedicated, disposable DBot Solana wallet (SolTest). NEVER the Maestro-shared execution wallet. Empty REFUSES the sync on Solana rather than falling back to anything

DBot: EVM execution wallet address — The dedicated, disposable DBot EVM wallet (BaseTest). NEVER the Maestro-shared execution wallet: DBot's risk posture (#2512) is fresh wallets holding float only. Empty REFUSES the sync for EVM chains rather than falling back to anything

DBot: max market cap (USD) — Only copy buys into tokens below this market cap. 25000 is the band the operator actually trades; it lives here rather than in a constant because that band moves (#1740). 0 = no ceiling, which reaches DBot as null rather than as 0 (#1620). Maps to buySettings.maxTokenMCUSD

DBot: buy slippage (%) — Tolerance the DBot copy engine buys with. Separate from ct_maestro_buy_slippage (default 100, max 1000) because DBot expresses slippage as a FRACTION in (0,1] and Maestro as a percent that legitimately exceeds 100: the same stored number cannot mean both, and reusing it would push "unlimited" on every DBot buy. Maps to buySettings.maxSlippage as pct/100

DBot: only one buy per token — #1464 in DBot's vocabulary. The copy signal is the lead's FIRST entry into a token; a re-entry is a different signal at a different price. On by default because that rule is not optional on this lane — it is the one the sim book already enforces. Maps to buySettings.buyOncePerWallet

DBot: refuse unsafe tokens at BUY time — On this lane nothing of ours runs before a buy: DBot copies the lead on its own and we only hear about the fill afterwards, so the task we push IS the entire pre-trade check. WHAT IT DELIVERS DEPENDS ON THE CHAIN: freeze authority and mint authority everywhere, plus GoPlus honeypot and unverified-source on EVM only — DBot documents those two as "valid for EVM", so on SOLANA THERE IS NO HONEYPOT SCREEN AT BUY TIME whatever this is set to. On by default because the alternative is no buy-time protection at all. Maps to buySettings.skipFreezableToken and skipMintableToken, and to skipHoneyPot and skipNotOpensource on EVM chains

DBot: sync cadence (minutes) — How often the runner READS BACK and re-verifies every dbot-lane wallet's copy task (#2560). Not a heartbeat: every pass compares the whole landed config against the goal, so this is also how quickly a task edited on DBot's own dashboard is corrected. The full-state rewrite only goes out when that comparison finds a difference (#2661) - an unchanged task is verified, not rewritten. Floored at 60s so a zero cannot busy-loop the price thread it rides on

DBot: copy tasks ARMED (#2560) — OFF means every task the sync writes is created and kept with `enabled: false`: DBot holds the exact config, echoes it back, and buys nothing. It stays off until the fill consumer (PR 3) and the sell signer by custody (PR 4) are merged, because a task that buys before those exist spends real money into a bag with no position row, no fill reader and no key that can sell it. Arming is an operator act from a browser session; the admin token cannot set it (#1756)

Maestro Sync Enabled — Auto-sync top-graded wallets to Maestro" (card note: "Sync approved wallets to Maestro on Telegram for MEV/same-block copy trade buys. Telegram API credentials must be in secre

⚠️ Allow Maestro Auto Buys — Master gate. When ON, Maestro spends real funds on detected buys (all chains). Leave OFF until you're ready to trade live." [copy_trade.html:1928-1929]

Buy % of the tracked wallet's buy — The percentage of the LEAD's buy Maestro replicates, clipped to Max Buy above — so Max Buy is only a CEILING until this is set high enough to exceed it. 0 = push n

Buy Slippage % (per wallet, when Smart Slippage off) — Pushed to each Live wallet so the per-wallet override matches your intent (the per-wallet value is what Maestro actually uses for buys). On chai

Smart Slippage — Auto-adjust for token taxes (per wallet)" [copy_trade.html:2112-2113]

Frontrun (BSC) — Buy before tracked wallet" [copy_trade.html:2120-2121]

Push Min Liq to Maestro's Auto Buy Checks — When on, every reconcile writes Minimum Liquidity (USD) — the value set in Liquidity Exit above, ct_min_liquidity_usd — into each LIVE wallet's Auto Buy Ch

Global Max MC ceiling (USD) — The market-cap ceiling for Maestro's GLOBAL Auto Buy Checks (per chain — wallet screens inherit it). 0 = CryptoBot does not own this; Maestro keeps its current value — n

Push Max Buy Tax % to Maestro — 0 = off (default): nothing is written, Maestro's Buy Tax check is left exactly as you set it. Above 0, every reconcile writes this ceiling into each LIVE wallet's Auto

Push Max Sell Tax % to Maestro — 0 = off (default), same as above but for the sell side — the honeypot-shaped one, since a token you can buy and cannot sell shows up here first. NOTE: setting it back

Read back after a sync — On (default, #874): after a successful add, read the wallet's Maestro settings page back and log what it actually holds. Costs one extra read-only Telegram round-trip per add

Auto-heal on drift — On (default): a reconcile that finds a LIVE wallet's Maestro state out of line re-pushes it, and reverts the wallet to SIM if it cannot. Off leaves the drift in place and only re

Reconcile every [n] hours — Sweep cadence. Default 6." [copy_trade.html:2092-2098]

Max buy — Base (ETH) — Native ETH ceiling written into each LIVE wallet on Base. Default 0.02." [copy_trade.html:2047-2050]

Max buy — BSC (BNB) — Native BNB ceiling on BSC. Default 0.08." [copy_trade.html:2053-2056]

Max buy — Solana (SOL) — Native SOL ceiling on Solana. Default 0.3." [copy_trade.html:2059-2062]

% of Balance Sizing — Nightly, set Max Buy to a % of the execution wallet's USD balance so buys compound as it grows (Base only)" [copy_trade.html:1945-1946]

Max Buy % (of balance) — When enabled, this overwrites Default Max Buy nightly (midnight ET). e.g. 1 = 1% of balance." [copy_trade.html:1952-1955]

Buy Monitor Enabled — Poll the watched Solana wallets for buys via Helius" (card note: "The watch set is automatic: Solana candidates in the ≥60 pool (top score first, max 25) plus active Solana copy

Monitor Poll Interval (seconds) — Helius credits: ~1/wallet/poll (≈8.6k credits/day/wallet at 10s) + ~100 per detected-buy history fetch" [copy_trade.html:2567-2570]

WS buy detector enabled — Off by default. On, a logsSubscribe socket watches the same wallet set the poller does, and detects a buy in the moment rather than on the next 90s tick. The poller keeps ru

WS detections are PRIMARY — On (default): a notification triggers the buy path immediately. Off is shadow — the socket detects and logs but never fires a buy, which is the half-step for watching the

Sim Copy-Trading (#1065) — Route detected Solana buys into SIM positions (token gate + $100 fib ladder). OFF = shadow mode, log only. No real funds either way." [copy_trade.html:2559-2560]

Solana selling ENABLED (#1559 rung 3) — The master switch for the Solana sell path. OFF = no sell is attempted at all. This is the control to reach for first if Solana sells are behaving badly — it s

Solana sells are DRY RUN — ON = the whole sell path runs against real quotes and stops before signing (rung 2). OFF = sells sign and broadcast for real (rung 3). Turning this ON is the reversible hal

Sell slippage (%) — Tolerance on an ordinary laddered sell. Default 10." [copy_trade.html:2615-2618]

Exit slippage (%) — The wider tolerance used when the goal is to be OUT — a stop, a rug exit, a full close. Deliberately far above the laddered figure: refusing to fill is the expensive outcome there

Priority fee (micro-lamports) — Per-transaction priority fee on the sell path. 0 = none (default). Raise it when sells are landing late in a congested block; it buys inclusion, not price." [copy_trad

Resolve Lookup Tables / v0 Routes (#1789) — OFF = legacy-only quotes, every account inline. ON = accept v0 routes and resolve their address lookup tables before the pre-sign check. Neither arms nor s

Enable the tiered daily Helius credit ceilings the RPC fetcher paces against (helius_health + solana_tx_fetcher own the tiers).

Intraday candidate pass (minutes) — Interval between intraday candidate passes. Default 180. The watermark saying when it last ran is candidate_intraday_last_run and is not a setting." [copy_trade.ht

PnL cap per candidate position (USD) — Ceiling applied to a single tracked position's contribution, so one implausible row cannot carry a wallet's score. Default 2500." [copy_trade.html:2877-2880]

Intraday score resnapshot (minutes) — 0 = off (default). When on (floored at 15m), today's wallet snapshots are rewritten on this interval by the funnel watchdog, so scores and grades move during the

Per-wallet ingest budget (seconds) — The same bound for the ingest phase (#1734), which the budget above deliberately did not cover. Sampling found the three longest holds in a 69-minute window all i

Per-wallet reprice budget (seconds) — Most wall-clock one wallet's reprice phase may take before the pass moves on (#2243). 0 = no bound. Default 600. A cut wallet keeps every position it already rep

Hot lane ends at (days) — Younger than this is HOT — the fib-active window." (card note: "One coherent trade: API spend against how stale a tracked candidate's price may get. ... Lowering these raise

Warm lane ends at (days) — Older than this is COLD. Between the two is WARM." [copy_trade.html:2802-2805]

Hot TTL (seconds) — 0 = every pass, which is the default and the point of the lane." [copy_trade.html:2811-2814]

Warm TTL — still moving (seconds) — Peak ≥ 1.2x. Default 21600 (6h)." [copy_trade.html:2819-2822]

Warm TTL — flat (seconds) — Default 43200 (12h)." [copy_trade.html:2825-2828]

Cold TTL — still a runner (seconds) — Peak ≥ 2x. Default 345600 (96h) — above the measured candidate cycle, or the skip can never fire (#2229)." [copy_trade.html:2831-2834]

Cold TTL — never ran (seconds) — The cheapest lane. Default 432000 (120h) — MAX_TRACK_DAYS is still its backstop (#2229)." [copy_trade.html:2837-2840]

GeckoTerminal rate limit (calls/min per endpoint) — 0 = built-in default (30/min, the vendor's documented keyless ceiling). Applied by the endpoint pool within ~5 minutes, no restart; clamped to 1–30

GeckoTerminal tracker reserve (calls/min per endpoint) — 0 = off, one shared budget (today's behaviour). Nonzero holds back this many of each endpoint's slots for the candidate-tracker pass so the wa

Rest these exits from the GeckoTerminal lane — Comma-separated endpoint names (as they appear in /api/copy-trade/gecko-stats). A rested exit keeps every other duty — RPC, WSS relay, health probe — an

Blockscout gateway daily credit ration — the self-imposed cap on keyed-gateway spend per UTC day, held against the DevPortal allowance. A SETTING rather than a constant because the right number turns on how that ledger renews, and reading it wrong once already cost a month of gateway: 3,000/day came from treating a 100k allowance as monthly, and the portal's renewal countdowns then showed it resets DAILY. 80,000 leaves 20% headroom on a 100k/day allowance. The day's spend is persisted per UTC day, so the ration binds across restarts (#2432).

Phase deadline hour (ET) — The hour by which the nightly phase must have completed. Default 6." [copy_trade.html:2935-2938]

Promote on REBASED PnL — Which PnL the promotion decision reads — a decision about real money, and until now one the operator could neither see nor set. On is the default. Turning it off puts the pas

Validation Alerts — Discord ping on first dry-run BUY and SELL for a Live-mode wallet" [copy_trade.html:1890-1891]

Session URL — Embedded in alerts so the operator knows where to paste log excerpts for review" [copy_trade.html:1897-1900]

Discord lane filter — Which lane's activity reaches Discord. both is the default." [copy_trade.html:2658-2665]