Real-time crypto market intelligence, right in your terminal.

HyperData Terminal streams public WebSocket feeds from Hyperliquid, Binance, Bybit, OKX and Deribit into six live dashboards, a local REST API and a paper-trading engine. Open source under Apache-2.0. No API keys to get started.

Install in four commands Read the source

Replay of a hyperdata session: the command starts, initializes its liquidation feed, order-flow engine, position scanner and market-data components, prints the six-dashboard menu, and opens the market overview with prices, change, funding and open interest for BTC, ETH, SOL and HYPE.

Exchanges: 5
Dashboards: 6
Data components: 14
API endpoints: 17+
License: Apache-2.0
Runtime: Python 3.12+

Six dashboards. One command.

Run hyperdata and pick a view from the menu. Every dashboard reads from the same data hub and updates in real time.

The six dashboards, what each shows, and where its data comes from
Key	Dashboard	Shows	Source
[1]	Liquidation Watch	BTC positions closest to liquidation on Hyperliquid. Distance-to-liquidation is tracked in real time, so you can see which whales are about to get wiped.	Hyperliquid
[2]	Liquidation Stream	Multi-exchange liquidation feed from Hyperliquid, Binance, Bybit and OKX, with size, price and exchange for every event.	4 exchanges
[3]	Liquidation Heatmap	Price-level view of where liquidations are concentrated for all cryptos: long-liquidation risk below price, short-liquidation risk above it.	4 exchanges
[4]	CVD Order Flow	Cumulative volume delta for BTC, combined from Hyperliquid and Binance trades, with each venue's share shown separately. See whether buyers or sellers are in control.	Hyperliquid + Binance
[5]	Market Overview	Funding rates, open interest and prices for 50 assets across exchanges. Spot divergences and funding extremes at a glance.	Cross-exchange
[6]	Whale Tracker	The largest open positions on Hyperliquid: size, entry price, PnL and liquidation price for the biggest players.	Hyperliquid
[0]	All Dashboards	The combined view, everything at once.	All

HyperData Terminal's combined view: liquidation watch, order flow, liquidation feed, market overview, HLP vault, system status, market intelligence, smart money and whales panels side by side in one terminal window. — The combined view. Each panel is one of the data components below, rendered by Rich in a single terminal window.

Five exchanges. Public feeds. No keys.

All data is fetched live from public exchange endpoints over WebSocket and REST. Nothing to sign up for, nothing to configure before the first run.

Inside, one hub owns 14 data components and manages their async lifecycles. The dashboards, the API and the paper-trading engine all read from that same hub, with no duplication.

Liquidations from 4 exchangesHyperliquid, Binance, Bybit, OKX
CVD order flowHyperliquid + Binance trades
Whale positionsHyperliquid
Funding ratesby exchange
Open interest50 assets
Orderbook depthBinance
DVOL implied volatilityDeribit
Long/short ratios
Spot prices
Smart-money signals

Ten of the fourteen components, as listed in the project README.

Data sources, what each provides, and how it is connected
Source	Data	Connection
Hyperliquid	Trades, positions, liquidations, funding, whale tracking	WebSocket + REST
Binance	Trades, liquidations, orderbook	WebSocket
Bybit	Liquidations	WebSocket
OKX	Liquidations	WebSocket
Deribit	DVOL implied volatility	REST

HyperData Terminal starting up: the HYPERDATA banner, the list of components being initialized (liquidation feed, order flow engine, position scanner, market data), the active dashboards, and a progress bar reporting all systems online. — Start-up. The hub brings each component online before the menu appears.

Build on the same hub.

Everything the dashboards see is available to your own code: through a local REST API and WebSocket, or directly inside a paper-trading strategy.

REST API and WebSocket

Start the server alongside or instead of the terminal. It binds to 127.0.0.1 by default and refuses a public bind unless you set an API key or explicitly accept the risk.

$ python run_api.py --port 8420

$ curl http://localhost:8420/v1/market/BTC | python -m json.tool
$ curl http://localhost:8420/v1/orderflow/BTC
$ curl http://localhost:8420/v1/liquidations/stats
$ curl http://localhost:8420/v1/positions/danger-zone

All 17 endpoints

REST and WebSocket endpoints
Endpoint	Returns
GET /v1/live	Minimal liveness probe, always unauthenticated
GET /v1/health	Server status and uptime
GET /v1/market	All assets: prices, OI, funding
GET /v1/market/{symbol}	Single asset detail
GET /v1/liquidations	Recent liquidation events
GET /v1/liquidations/stats	Aggregate liquidation statistics
GET /v1/orderflow/{symbol}	CVD snapshots for a symbol
GET /v1/funding-rates	Funding rates across exchanges
GET /v1/funding-rates/{symbol}	Single asset funding rate
GET /v1/long-short-ratio	Long/short ratio
GET /v1/basis	Basis spread
GET /v1/deribit/iv	DVOL implied volatility
GET /v1/orderbook/{symbol}	Orderbook snapshot
GET /v1/whales	Top whale positions
GET /v1/positions/danger-zone	Positions closest to liquidation
GET /v1/public/metrics	Server metrics and component health
WS /v1/ws	Real-time event stream: liquidations, trades, signals

Optional LLM agent

The agent sends a market summary to any OpenAI-compatible API and asks for a trading decision. Point it at OpenAI, Ollama, LM Studio, Groq, Together or your own endpoint with three variables in .env: LLM_BASE_URL, LLM_MODEL and LLM_API_KEY. The prompt lives in src/strategies/llm_agent.py.

Paper trading, your strategy

Real market data, fake money. A strategy is one class with one method that returns a signal or nothing. The whole interface fits in about 30 lines.

from src.strategies.base import Strategy, Signal

class MyStrategy(Strategy):
    name = "my_strategy"

    def evaluate(self, hub) -> Signal | None:
        # hub.orderflow, hub.market, hub.liquidations,
        # hub.funding, hub.lsr, hub.positions,
        # hub.spot, hub.deribit are all live here.
        snap = hub.orderflow.get_snapshot("BTC", "5m")
        if not snap:
            return None

        cvd = snap.buy_volume - snap.sell_volume
        if cvd > 100_000:
            return Signal("BTC", "BUY", size_usd=100,
                          confidence=0.7,
                          reason="Strong buy pressure")
        elif cvd < -100_000:
            return Signal("BTC", "SELL", size_usd=100,
                          confidence=0.7,
                          reason="Strong sell pressure")
        return None

Included example strategies

Example strategies shipped with the project
Strategy	Reads	Logic
CVDMomentum	hub.orderflow	Buy when order flow is strongly positive, sell when negative
FundingRateArb	hub.funding	Short when funding is high, long when negative
LiquidationCascade	hub.liquidations	Buy the dip after large liquidation cascades
WhaleFollow	hub.positions	Mirror the direction of the largest whale positions

Data that checks itself.

The hub continuously cross-checks its live feeds against Binance and Deribit and flags anything stale or drifting. The state is printed on the dashboard and exposed at /v1/health, so your own tools can read it too.

Coverage is honest, not a complete census. OKX and Bybit liquidations are real feeds, with Bybit covering its top 15 symbols. Binance's stream is throttled at the source to about one liquidation per symbol per second. Hyperliquid liquidations are inferred from large trades and flagged as estimated. Details are in docs/DATA_INTEGRITY.md.

Feed status states shown on the dashboard
State	Meaning
LIVE	The feed agrees with the reference. Data is arriving and matches the cross-check within tolerance.
STALE	The feed has gone quiet. No fresh data has arrived recently, so what you see may be out of date.
DRIFT	The feed disagrees with the reference. Values have diverged from Binance or Deribit beyond the expected range.

Four commands to live data.

Clone, install, run. The first launch opens the dashboard menu with data streaming from all five exchanges.

Requires: Python 3.12 or newer, pip, git
API keys: None for the dashboards or the API. Only the optional LLM agent needs one.
License: Apache-2.0

$ git clone https://github.com/Co-Messi/HyperData-Terminal.git
$ cd HyperData-Terminal
$ pip install -e .
$ hyperdata