Solana LP Agent

Live simulation controls

Start the live simulated bot. It stays 100% cash until Start is pressed, the fresh-pool gate is met, and the countdown expires.

Simulation not started yet.

Bot strategies recap

Live-only recap for Conservative / Balanced / Aggressive paper portfolios. Strategy lab/backtests are separated so they do not touch live state.

α vs hold

Loading bot state…

Allocation timeline

Live bot allocation/cash/cost timeline. Strategy Lab can run historical simulations without overwriting this live state.

Strategy Timeline point

Start the live simulation to load the live allocation timeline.

Research terminal

Pool Intelligence

Rank pools by fee density, stability, liquidity quality and execution risk. This page is read-only and does not touch live state.

Pool ranking

Higher health means better score, stronger fee density and fewer anomaly/liquidity penalties.

Safety layer

Risk & Anomaly Center

Monitor invalid prices, fallback rows, liquidity gaps and pools that should not be trusted by the strategy engine.

Risk events

Actionable warnings generated from current pools and DB status.

Strategy Lab · isolated backtest sandbox

Build strategy scenarios and compare them on saved DB snapshots. Lab tests never start, reset, buy, rebalance, pause or overwrite the live bot.

Strategy setup

These are strategy/risk/allocation parameters only. Initial capital, lookback window and fallback execution costs are common simulation inputs below.

Scenario Scenario name Base profileLive presets copy current live values and lock strategy fields. Choose Custom to edit values without touching live. Rank score profileChoose which saved Score Lab rank is used to filter/sort candidates in this Strategy Lab scenario. Allocation mode Max pools Reserve % cash Min score Max allocation cap % / pool Hard exit min score Hard exit max 1h move % Hard exit max IL bps

Live preset scenarios remain read-only here. Switch Base profile to Custom to unlock and tune the strategy. Saving/running here does not modify live strategy settings.

Simulation setup

Common inputs used for every selected strategy in this lab run. These do not modify live simulation settings.

Snapshot archive: loading…

Initial capital USDCommon starting cash for every strategy being compared. Lookback days Rebalance hours Check interval minutesCommon scan/check bucket used by all selected scenarios. Fallback gas USD/actionUsed only when snapshot/API gas cost is missing. Fallback swap/route bpsUsed only when snapshot/API route cost is missing.

Strategies included in simulation set

Choose strategies and run a simulation set to calculate results from collected data.

Scenario recap

Same compact summary format as live, showing only the lab scenarios included/run in this sandbox. The dashboard switch can show standard USD return or α vs 50/50 hold.

Please run first simulation.

Lab allocation timeline

Allocation/cash/cost timeline for the selected scenario. It does not touch live bot state.

Scenario result Timeline point

Please run first simulation to load the lab timeline.

Best candidate—

Visible pools—

Highest vol/liquidity—

Market mode—

Pool score history

Interactive chart from server-side snapshots. Bands follow your ENTER/HOLD/WATCH/AVOID thresholds.

Pool Settings

Choose what pools are tracked on the Dashboard. Discovery helps find candidates; tracking makes them active.

Pool discovery

Controls the discovery table below. Dashboard only shows pools you explicitly Track and keep active.

Discovery enabled Discovery mode Discovery search query Discovery default limit Min 24h volume USD Min liquidity USD Min txns 24h

Tracked custom pools

Only active tracked pools appear on the Dashboard. Use Track below to add pools from discovery.

Auto top pools discovery

Discovery only: nothing appears on the Dashboard until you click Track. Columns are sortable and include score, estimated fee APR, IL24h and fee tier used.

Search queryLimit

Add / update pool

Use exact symbols to avoid BNSOL/SOL-style mistakes. Example: base SOL,WSOL, quote USDC.

ID optionalLabelDEX Screener search queryBase symbols comma separatedQuote symbols comma separatedTypeFee rate decimalCorrelated pair?

Pool Rating Settings

Configure the pool rating engine: LP fee estimate model, score formula and action thresholds. These settings rate pools; they do not control live allocation sizing.

LP fee estimate model

This section is only about estimated LP fees earned by providing liquidity — not gas, swap route fees or execution costs. Instant uses only current 24h volume. Rolling averages saved snapshots over N days. Weighted gives more importance to recent data.

Fee mode

Rolling days: 7

Weight last 24hWeight recent 2–7dWeight older

Fee APR formulafeeAprPct = (smoothedVolume24h / liquidityUsd) × feeRate × 365 × 100

Action thresholds

These decide the label shown on each pool. Keep them ordered: Enter > Hold > Watch.

Enter candidate ≥Hold ≥Watch ≥

Editable scoring formula

The score is always capped at 100. Positive sliders rebalance to exactly 100.0/100.

AMM base rulex × y = k

For a classic constant-product pool, x is token A, y is token B, and k stays constant during swaps.

Impermanent loss estimater = 1 + change24h / 100
IL24h = | 2 × √r / (1 + r) − 1 | × 100

This compares AMM value after price moves versus simply holding the original tokens. It does not include earned fees.

Main score formula

score = clamp(base + feePotential + liquidityQuality + activityQuality + correlationBonus − trendPenalty − ilPenalty − ghostPenalty − lowFeePenalty, 0, 100)

Positive score budget = 100 max points

Base score —Fee potential max —Liquidity quality max —Activity quality max —Correlation bonus max —

Scaling and penalties

Fee APR target % Trend weight 1h Trend weight 6h Trend weight 24h Correlated trend multiplier IL penalty weight Correlated IL penalty weight Ghost pool penalty Low fee penalty Low fee APR threshold %

Live strategy settings

These settings affect the live simulated bot only from the moment you save onward. Saving does not reset cash, positions, runtime, PnL or history.

Debug panelsInitial capital USDGlobal check interval

—

Controlled in DB Management.Rebalance every hoursFirst buy delay minutes Min fresh pools before Start Backtest lookback daysGas USD per action Entry/exit swap cost bps Live cost API estimates Free APIs used: Solana public RPC for gas/priority fees + Jupiter quote API for route impact. LP pool fee tier is added separately. Fallbacks below are used if API fails.Fallback gas USD/actionFallback entry/exit bps

Hard exit conditions

The server checks tracked pools every X minutes. Hard exits can force an immediate simulated exit on the next scan.

Hard exit enabled Hard exit if score belowHard exit if 1h move ≥ %Hard exit if IL24h ≥ bpsHard exit if score drops by

Strategy risk profiles

Allocation mode

Caps are used only when Allocation mode is Dynamic with cap. Equal mode ignores score weighting after pool selection and splits deployed capital evenly across selected pools.

Conservative

Max poolsReserve % cashMin scoreMax allocation cap % / single pool

Balanced

Max poolsReserve % cashMin scoreMax allocation cap % / single pool

Aggressive

Max poolsReserve % cashMin scoreMax allocation cap % / single pool

How the simulation works

Allocation rulepick top pools where score ≥ minScore, capped by maxPools; keep reserve % in cash; weights follow allocation modeLP fee estimate per intervalfeePnl ≈ allocatedUsd × (feeAprPct / 100) × (hours / 24 / 365)Directional LP market exposurelpValueRatio ≈ √(priceNow / pricePrevious)

The server checks tracked pools every configured X minutes and saves snapshots. The simulated bot can hard-exit immediately on any scan if risk thresholds are hit. New entries and normal rotations happen on the rebalance cadence.

Scoring research

Score Lab

Optimizer for how pools should be scored before strategy thresholds are applied. Analysis is read-only; saved score profiles are persisted in the DB and can be selected from Strategy Lab.

Analysis window Universe

ModeAnalysis read-only ProfilesDB persisted Live stateNot affected

Metric dictionary & formulas

Every metric used by Score Lab must be explainable: what it measures, where it comes from, and why it may matter.

Formula analysis

Tests fee density, volume/liquidity, liquidity depth, transaction activity, price/IL stability, score momentum, cost efficiency and DEX risk across selectable time windows.

Ready to analyze current DB. Select an analysis window and click Analyze score formula.

Score profile builder

Flag the metrics to include, let Score Lab propose weights, manually tune them, then save a named score profile. Strategy Lab can then use that profile as the candidate rank.

Saved score profiles

Profiles are rank formulas saved in the DB, not live trades. They are available after refresh and from another device using the same backend DB.

API & cost settings

Control live fee/gas estimates used by paper buy, rebalance and hard exit. If an API fails, the bot uses the fallback values below and stores the source as fallback.

Gas APISolana RPC estimatestatus -

Swap/route APIJupiter quotestatus -

Pool searchDexScreener searchstatus -

Pool token-pairsDexScreener token-pairsstatus -

Pool pairsDexScreener pairsstatus -

Chart APIQuickChartstatus -

Fallbacks- / -gas/action - swap bps

Cost APIs

Live cost API estimatesON = try real Solana RPC gas/priority fee + Jupiter quote first. If they fail, use fallback values and mark source=fallback. Solana RPC URL status -Used for health check and gas estimate via getRecentPrioritizationFees. Jupiter quote API URL status -Used to estimate route price impact/slippage bps. Fallback gas USD/action Default: $0.05 per action. Fallback swap/route cost bps Default: 8 bps per entry/exit route. Legacy gas USD/action Kept synced for old rows/settings. Legacy entry/exit cost bps Kept synced for old rows/settings.

Pool/data APIs

DexScreener search API status -Used by Auto top pools discovery. DexScreener token-pairs API status -Used for token-pair lookups. DexScreener pairs API status -Used for pair-address lookups. Telegram API fixed

https://api.telegram.org

Uses server env token/chat ID; URL fixed for safety. QuickChart API status -

https://quickchart.io/chart

Only used when Telegram chart images are enabled.

Load settings to see current API/cost configuration.

How costs are applied

Buy / rebalancenetPositionUsd = allocatedUsd - gasUsd - routeCostUsd - LPFeeTierCostUsd Hard exitcashReceived = positionValueUsd - gasUsd - routeCostUsd - LPFeeTierCostUsd AccountingtotalValue = cash + deployedPositionValue; costs are already deducted and also shown as counters

Older simulation points are not retroactive. Run a new Buy / rebalance after deploying this version to see fresh gas/swap counters.

Telegram bot options

Automatic trade updates fire whenever a strategy enters/exits during a market scan. Reports are separate scheduled summaries.

🤖 Master enabled 🔐 Bot token

Server env: not configured

Set TELEGRAM_BOT_TOKEN in Plesk environment variables. 💬 Chat ID

Server env: not configured

Set TELEGRAM_CHAT_ID in Plesk environment variables. 🚀 Start alerts 🕒 Scheduled reports Report every hours ⚡ Automatic trade alerts 🟢 Notify entries 🔴 Notify exits 🧠 Include reasons 🥧 Include allocation after trade 📈 Attach performance chart 🥧 Attach allocation pies

Chart images use an external chart image URL and are optional. Keep off if your hosting blocks external requests.

Security note: Telegram secrets are read only from server environment variables and are not stored in the DB or exposed to the frontend.

DB Management

Control the local server DB. Default mode should be Continue DB universe: after the first scan, the bot keeps collecting the same discovery pools after restart.

Data Management

Separate raw snapshot archive from replayable strategy logs. Snapshots are the market-data source of truth; strategy logs can be exported or cleared without touching snapshots.

Snapshot Archive

Export raw data, export cleaned data with invalid price rows flagged/fallback-ready, or import a snapshot archive for deterministic reruns.

Export raw snapshots Export cleaned snapshots

Append instead of replace

Strategy Logs

Strategy logs are disposable replay/debug outputs. Clearing them does not delete market snapshots.

Export strategy logs Export split DB

Persist cleaned DB

After validate/clean/replay, make the cleaned snapshot archive the default JSON DB. Semantic timestamped backups are created for previous default and cleaned snapshots, so RAW/CLEANED/REPLAYED states stay traceable.

Backups will appear below with restore/delete controls.

Replay / Rebuild

Rebuild the live paper recap from cleaned snapshots using the current engine. This preserves raw snapshots, clears contaminated strategy logs, and replaces only replayable live portfolio state.

Replay lookback

Discovery universe

Continue DB universe = keep using the discovery pools already saved in DB. Best for consistent ML data. Rescan every scan = always refresh top pools from API; useful only if you want the universe to change continuously.

Discovery mode Rescan limit Sampling rate minutes

Global check interval. This value controls how often the server scans the market, writes snapshots to the DB, evaluates simulation/strategy decisions, and triggers Telegram trade alerts. It overwrites the old Simulation Settings check interval because the whole system uses the same scan loop.

Cleanup tools

Use these only when you intentionally want to reduce DB size or restart the simulation. Snapshots are the valuable AI dataset, so do not clear them casually.

Keep latest days

Delete data for one saved pool

Use this if one pair polluted the dataset. You can delete only its snapshots while keeping the pool config, or delete snapshots and remove it from the discovery/custom universe.

Saved pool / pair Also remove pool config

DB Visualizer

Interactive admin browser for snapshots, pools, settings and simulation state.

Table Search Limit Range Pool filter Sort DEX/source filter Min liquidity USD Min volume 24h USD Min score

Click a row to inspect/edit JSON. Save replaces that DB row; Delete removes it after confirmation. Cost fields show API vs fallback gas/swap estimates when available.

Choose a table and refresh.

Rows

Compact table view.

Selected row editor

Raw record for verification/debug. Advanced: edit carefully.

LP opportunity dashboard

Recent API requests

Screen call flow

Live simulation controls

Bot strategies recap

Allocation timeline

Pool Intelligence

Pool ranking

Risk & Anomaly Center

Risk events

Strategy Lab · isolated backtest sandbox

Strategy setup

Simulation setup

Strategies included in simulation set

Scenario recap

Lab allocation timeline

Pool score history

Pool Settings

Pool discovery

Tracked custom pools

Auto top pools discovery

Add / update pool

Pool Rating Settings

LP fee estimate model

Action thresholds

Editable scoring formula

Positive score budget = 100 max points

Scaling and penalties

Live strategy settings

Hard exit conditions

Strategy risk profiles

Conservative

Balanced

Aggressive

How the simulation works

Score Lab

Metric dictionary & formulas

Formula analysis

Score profile builder

Saved score profiles

API & cost settings

Cost APIs

Pool/data APIs

How costs are applied

Telegram bot options

DB Management

Data Management

Snapshot Archive

Strategy Logs

Persist cleaned DB

Replay / Rebuild

Discovery universe

Cleanup tools

Delete data for one saved pool

DB Visualizer

Rows

Selected row editor