Live Every symbol scored once a minute

Know which perp markets are loaded to move.

StoLiq collects the data that precedes a pump or a dump — open interest, funding, liquidations, order book depth and on-chain whale flow — and ranks every symbol so the highest-risk markets surface first.

StoLiq ranks conditions, not outcomes. A high score means a market is positioned to move violently — not that it will go up. Not financial advice.

StoLiq — Top risk
  Updated 12s ago · illustrative data

Top risk

#SymbolScoreLeanOI 1hFunding
1RAVEUSDT
90
Dump+38.4%−0.31%
2NOVAUSDT
61
Dump+21.0%+0.18%
3KAIUSDT
52
Pump+12.7%−0.09%
4ORBUSDT
47
Dump−9.3%+0.12%
5VEXUSDT
43
Pump+6.1%−0.04%
6ZENUSDT
39
Dump+3.8%+0.07%
RAVEUSDT · SCORE ≥ 70
hourly, last 7 days
SCORE BREAKDOWN
RAVEUSDT
Dump-lean
90/ 100 · relative rank, not a probability
holder_concentration
14
oi_surge
22
thin_book
10
funding_extreme
16
exchange_inflow
8
volume_divergence
10
price_momentum
12
rsi_extreme
8
liquidation_cascade
14
new_listing
6
Binance perpetuals Bybit perpetuals Ethereum BSC Solana CoinGecko metadata TimescaleDB
1 minScore cadence, every symbol
10Transparent score components
2Exchanges: Binance & Bybit
3Chains: ETH, BSC & Solana
Features

One ranking. Every signal behind it, on the record.

Violent moves leave footprints in derivatives, order books and wallets long before they show on a chart. StoLiq stitches them together and keeps the receipts.

Derivatives pressure

Funding, open interest, liquidations, 1-minute klines and long/short ratio. Funding is measured as a z-score against the symbol’s own 30-day history, not a one-size threshold.

FundingOpen interestLiquidations

Order book depth

Snapshots aggregated into depth, spread and imbalance. High-scoring symbols are polled more often, and open interest is judged against the depth that has to absorb it.

DepthSpreadImbalance

On-chain whale flow

Holder concentration and whale transfers to and from labelled exchange wallets, so an insider deposit landing on an exchange is a number, not a rumour.

Top-10 holdersExchange inflow

Every alert is auditable

Each score is stored with its full component breakdown and the feature snapshot it was computed from, so any alert can be explained after the fact.

Component breakdownFeature snapshot

Tunable, without a redeploy

Component weights live in the database and are re-read every minute. Retune sensitivity through the API and the next score reflects it.

Runtime weightsToken priority

Honest about missing data

Missing data can only pull a score down. A symbol with almost no coverage can’t look dangerous just because the little data it has is extreme.

Conservative by design
How it works

From raw ticks to a ranked board in four steps.

Collectors own their WebSocket subscriptions and block cursors, so they run as singletons. Only the read API scales out.

01 · COLLECT

Collectors

Derivatives, order book and on-chain collectors stream and poll continuously. A token mapper links exchange tickers to CoinGecko metadata and contracts.

derivativesorderbookon-chain
02 · STORE

TimescaleDB

Hypertables and continuous aggregates hold the history. Compression and retention policies keep the database from filling up.

hypertablesaggregates
03 · SCORE

Signal engine

Features become ten components, each with an intensity from 0 to 1 and a pump/dump split. The weighted sum, normalised by total weight, is the score.

pure functionsevery minute
04 · SURFACE

Board & API

The highest-risk symbols rise to the top of the board and the read API, with the full breakdown one click away.

top-risksymbol detail
The reference case · RAVEUSDT

A low-float token, a thin book, and an insider deposit.

Every one of these is a component of the score — and that combination is exactly what the model is tuned to rank first. It is also pinned by tests: the RAVEUSDT signature must rank far above both ordinary volatility and an equally violent selloff.

~93%Top-10 wallets’ share of supply

Holder concentration at an extreme, and rising.

OI ↑↑Open interest exploding

Against an order book too thin to absorb it.

Funding ↓Deeply negative

Because the crowd was positioned short.

HoursInsider deposit to exchange

Landing before the vertical move.

Score model

Ten components. Published weights. No black box.

Scoring is pure functions: no database, no clock, no network. The same inputs always produce the same score, so every number can be reproduced.

  • Relative, not probabilistic. Scores rank symbols against each other; they are not odds of a move.
  • Direction-aware. Each component declares how its intensity splits between pump and dump.
  • Retunable. Weights are stored in the database and re-read once a minute.
ComponentDefault weightReads
oi_surge
22
Open interest change over 1h and 24h
funding_extreme
16
Funding z-score against the symbol’s own 30-day history
liquidation_cascade
14
Hourly liquidations versus a 7-day baseline, split by side
holder_concentration
14
Top-10 holder share and its change
price_momentum
12
Price change over 1h and 24h
volume_divergence
10
Open interest versus turnover; price rising while spot volume fades
thin_book
10
Open interest relative to visible depth
exchange_inflow
8
Whale deposits into exchange wallets, absolute and versus market cap
rsi_extreme
8
1h RSI at either extreme
new_listing
6
Age of the listing
API

Built to be read by machines, and audited by people.

A read API for rankings and symbol detail, plus runtime tuning of the scoring weights. Retune sensitivity or move a token to the front of the polling queue with one request.

  • Pipeline health, per source. Reports how long ago each source last wrote a row — catching the failure liveness probes miss: a process that is up and silently writing nothing.
  • Watch one token closely. Raise a token’s priority to front-load on-chain and order book polling.
  • Interactive docs. API docs are served by the API itself at /docs.
# Per-source freshness and overall pipeline status
curl -s "$STOLIQ_API/api/v1/health"
Pricing

Subscribe for the speed you need.

Start on the free board. Upgrade for real-time scores, alerts and the API. Switch plans whenever your desk changes.

Save 20%

Free

Look around the market’s risk board with a 15-minute delay.

$0 forever  
  • Top-10 risk board, 15-min delay
  • Binance perpetuals
  • Total score only
  • 3 watchlist symbols
  • 24-hour history
  • Community support

Desk

For small desks and power users who want to tune the model.

$99 / month Billed monthly
  • Everything in Pro, plus:
  • Retune component weights via API
  • Token priority for faster polling
  • 500 watchlist symbols
  • Webhook alerts
  • REST API · 600 req/min
  • 1-year history
  • 5 seats

Enterprise

Run it yourself or let us — with your exchanges and chains.

Custom Annual contract
  • Everything in Desk, plus:
  • Self-hosted (Docker / Kubernetes)
  • Additional exchanges & chains
  • Your own exchange-wallet labels
  • Unlimited seats & symbols
  • Dedicated support

Compare plans

FreeProDeskEnterprise
Data
Score refresh15 min delay1 min1 min
ExchangesBinanceBinance, BybitCustom
Symbols visibleTop 10All trackedAll tracked
Component breakdown
History24 hours1 yearCustom
Alerts & API
Watchlist3500Unlimited
Telegram, Discord & email alerts
Webhook alerts
API rate limit600 / minCustom
Control
Retune score weights
Token polling priority
Self-hosted deployment
Seats15Unlimited
SupportCommunityPriorityDedicated
FAQ

Questions, answered plainly.

Does StoLiq tell me when to buy or sell?
No. StoLiq is not a trading system and not a predictor. It ranks conditions: a high score means a market is positioned to move violently, not that it will go up. The numbers are relative rankings, not probabilities.
What does the score actually measure?
Ten components — open interest surge, funding extremes, liquidation cascades, holder concentration, price momentum, volume divergence, thin books, exchange inflows, RSI extremes and new listings — each map a family of features onto an intensity from 0 to 1. The weighted sum, normalised by total weight, is the score.
What if data is missing for a symbol?
Missing data can only ever pull a score down. A symbol with almost no coverage can’t look dangerous just because the little data it has is extreme. Pipeline health is also reported per source, so a silently idle collector is visible.
Which exchanges and chains are covered?
Binance and Bybit perpetuals for derivatives and order book data, and Ethereum, BSC and Solana for on-chain holder and whale-transfer data. Enterprise plans can add exchanges and chains.
Can I retune the model?
Yes, on Desk and above. Component weights are stored in the database and re-read once a minute, so changes made through the API take effect without a redeploy. You can reset to defaults at any time.
Can I run StoLiq on my own infrastructure?
Enterprise includes a self-hosted deployment. Every service is the same image with a different entrypoint, so it runs the same way on a laptop with Docker as it does on Kubernetes. Collectors and the signal engine are singletons by design; only the API scales horizontally.
Can I switch plans or cancel?
You can move between plans at any time from your account. Annual billing is discounted by 20% compared with paying monthly.
Get started

See what’s loaded to move before it moves.

Start on the free board in under a minute. Upgrade when you want real-time scores and alerts.