How Hyperliquid Market Data Is Structured

Intermediate | 3 min | Last verified 2026-09-07

Hyperliquid market data is split across metadata, order books, trades, candles, prices, funding, open interest, and account-specific state; each dataset answers a different question.

Hyperliquid does not have one universal “market data” feed. Market metadata tells you what can trade. L2 books show resting liquidity. Trade streams show executions. Mark and oracle prices support risk. Funding and open interest describe derivatives positioning. Candles aggregate past trades. A research or trading system should keep those objects distinct instead of collapsing them into one price series.

Metadata defines the market

Before interpreting a number, identify the asset index, symbol, szDecimals, quote asset, leverage/margin settings, and market type. Spot and perpetual identifiers use different conventions in the trading API.

Metadata changes more slowly than a book but is just as important for correct unit conversion.

Order books show intention

L2 data reports resting bids and asks at price levels. It tells you the liquidity currently displayed, not what will definitely remain when your order arrives.

Book snapshots are essential for spread, depth, imbalance, and expected market-order impact.

Trades show executed interaction

Trade data records transactions that actually matched. Volume and aggressor-side analysis come from this stream, but executed trades cannot reveal hidden intent that never reached the book or orders that were cancelled.

Reference prices are risk data

Oracle and mark prices have roles distinct from the last trade. Hyperliquid uses reference-price systems for funding, margin, liquidations, PnL, and triggers. A market-research table should label each price explicitly rather than call all of them “price.”

Derivatives state adds another dimension

Funding, open interest, margin parameters, and liquidations describe leverage and positioning. A market can have rising price with falling open interest, rising price with aggressive open-interest expansion, or flat price with expensive funding. Those are different regimes.

Build an event model

For serious analysis, persist timestamps and source type for every event. Do not resample everything immediately into candles. The raw book/trade/funding/account events are what let you later ask better questions about liquidity and execution.