Get Candles Batch

Trailing OHLCV candles for up to 25 coins in ONE call — the same bars /hyperliquid/candles serves per coin, keyed by coin under series.

Trailing limit only; for an explicit start/end range use the single-coin route. Every coin must be in HL's universe (400 unknown_coin otherwise — nothing is silently dropped). On a cold cache a series that has not arrived within ~8s is listed in pending instead of holding the whole response, and retry_after_s says how long to wait (scaled to how many are still fetching): their fetches keep running and are cached, so retry just the pending coins after that. A cold coin costs Hyperliquid rate limit (~20 weight, whatever limit is), so a first pull of many coins takes a while and repeat pulls are instant; for a whole-universe scan that needs no candles at all, use /indicators/heatmap. atr and forming_bar_timestamp behave as on /hyperliquid/candles.

The contract: complete is true iff every requested coin is in series; a coin is never in both series and pending. wait (0-25s, default 8) holds longer for a caller that prefers one call over a poll. closed_only=true returns the last limit COMPLETED bars — the read for a strategy that acts on a closed bar — and is served from any series fetched since the latest close, so the liquid coins pre-warmed just after each 4h close stay instant for the whole bar. Fees for sizing are on /hyperliquid/meta (fees); funding + OI on the same bar grid are /backtesting/hl-funding-bars (join on timestamp).