Hyperliquid - HyperCore historical coverage starts at .Every data type (fill, order, trade, funding, and ledger event) from that block onward is retained and queryable.
Drop-in Hyperliquid API replacement
Drop-in replacement for both the public
/info REST API and /ws WebSocket. Change one URL to remove rate limits and unlock more data. Hosted at hypercore.goldrushdata.com.Real-time Streaming
Wallet firehose, OHLCV for every HIP-3 and HIP-4 market the moment it goes live, and pre-decoded liquidations and vault events.
HyperEVM coverage
Token balances, transfers, approvals, NFTs, and gas for HyperEVM - same shape as every other EVM chain on GoldRush.
Drop-in coverage: REST + WebSocket
GoldRush replaces both of Hyperliquid’s public surfaces with a single host -hypercore.goldrushdata.com. Same request bodies, same response shapes, same subscription payloads as the public API. Swap the URL, add the GoldRush API key, and your existing clients work unchanged - with no rate limits and no per-IP subscription caps.
Auth differs between REST and WebSocket. REST takes an
Authorization: Bearer <GOLDRUSH_API_KEY> header. The WebSocket takes a ?key=<GOLDRUSH_API_KEY> query parameter (most WebSocket clients can’t send headers during the upgrade handshake). Same key for both.REST /info API
Drop-in replacement for
POST https://api.hyperliquid.xyz/info. 17 wire-compatible type values plus GoldRush-native batch endpoints (batchClearinghouseState, batchSpotClearinghouseState) for up to 50 wallets per call.WebSocket API
Drop-in replacement for
wss://api.hyperliquid.xyz/ws with no 1000-subscription-per-IP cap. Subscribe to l2Book (aggregated snapshots - omit coin to stream every asset on one subscription), the GoldRush-native l2BookDiff (snapshot plus per-block diffs of changed {px, sz, n} levels - accepts a single coin, a list, or wildcard), or the GoldRush-native l4Book (per-order snapshot plus per-block diffs with user, oid, cloid, and trigger metadata exposed).Hyperliquid API limitations addressed by GoldRush
The public Hyperliquid/info API is generous, but it has hard limits that most production apps run into:
- 1200 weight/min/IP rate limits per address.
- 1000 WebSocket subscriptions per IP - not enough to track every active trader.
- WebSocket subscription filters are required - every subscription must specify an asset (e.g.
coinonl2Book), forcing per-asset fan-out that burns through the 1000-subscription cap. GoldRush makes these filters optional so a single wildcard subscription streams every asset. - No batch address endpoints - account-state calls are single-wallet only; GoldRush adds
batchClearinghouseStateandbatchSpotClearinghouseStatefor up to 50 wallets per request. userFillscapped at ~10,000 rows - active traders blow through that in weeks.- Liquidations buried inside fills as a thin stub; vault, staking, and delegation data arrives untyped.
- HIP-3 and HIP-4 discovery is manual -
candleSnapshotis poll-based and effectively limited to mainstream markets. - No HyperEVM concept at all - DEXes, NFT mints, and lending live on chain
999and the/infoAPI doesn’t see them.
What’s included
Infrastructure
GoldRush runs its own Hyperliquid indexing infrastructure - purpose-built to be the fastest, most reliable Hyperliquid data provider for production apps and quant desks:- Dedicated Hyperliquid nodes in Tokyo - co-located with Hyperliquid’s validator infrastructure for low-latency reads.
- Private backbone for high-throughput data ingestion.
- No rate limits - bypass the public
/infoand WebSocket constraints entirely. - Full historical backfill - every fill, funding payment, and ledger event back to HyperCore . HyperEVM coverage goes back to genesis.
Quickstart
Here are three “first 5 minutes” quickstarts. Pick whichever maps to what you are building.Upgrade from the public Hyperliquid API
Already using
api.hyperliquid.xyz/info? Change the URL and add one header.Watch wallets
Subscribe to thousands of HyperCore wallets in one connection.
Pipe orderbook fills into your warehouse
Stream
hl_fills into ClickHouse, BigQuery, or Postgres.