HyperCore is now live on Alchemy. Any team building on Hyperliquid can stream HyperCore order books, best bid and offer (BBO), mid prices, and fills over WebSocket, on the same API key they already use with us. Streaming is available for all Alchemy users and free during launch week.
Hyperliquid market data, without running a Hyperliquid node.
We opened HyperCore streaming in early access in October and worked through it with a first group of teams. Today it's open to everyone building on Hyperliquid.
What can you stream from HyperCore?
HyperCore is Hyperliquid's native trading engine, where orders, fills, and liquidations live directly in chain state. Nine WebSocket streams are live today:
Stream
What it delivers
l2Book
Aggregated price levels per market, as a full snapshot
l2BookDiff
A snapshot, then only the price levels that changed
bbo
Best bid and offer per market
allMids
Mid prices across markets
l4BookUpdates
Each individual order as it's placed, resized, or removed, with its place in the queue
allFills
Fills across the markets you choose
userFills
Fills for specific accounts
builderFills
Fills routed through specific builder codes
liquidationFills
Every liquidation fill
If you're new to how these records relate, our guide to Hyperliquid orders vs trades vs fills walks through each one and how they join.
On Enterprise, a single subscription can cover up to 200 markets or 200 accounts, and leaving out the market filter subscribes to every market at once. Every data message also carries the block number and block time it came from, so book changes and fills line up against the same chain clock.
Paid plans can also stream raw HyperCore blocks over gRPC.
How is this different from Hyperliquid's public API?
Hyperliquid's public API is free and needs no account, and it's where most teams build their first version. It's also a shared endpoint with per-IP limits, including 10 WebSocket connections and 10 unique users across user-specific subscriptions. Its order book feed sends one market per subscription, as a full snapshot pushed at most every half second.
HyperCore streaming on Alchemy changes four things for a trading system:
- Higher limits. The highest HyperCore rate limits in the industry, with up to 200 markets or accounts per subscription on Enterprise, compared with one on the public API.
- Changes instead of snapshots. l2BookDiff sends only the levels that moved, and l4BookUpdates adds order-level detail the public feed doesn't stream.
- Detectable gaps and fill replay. Each subscription numbers its messages, so a skipped number means a lost message. Fill streams accept a fromBlock and replay roughly the last minute of blocks after a reconnect.
- One provider for both layers. HyperCore runs on the same account and API key as your HyperEVM RPC, served from Hyperliquid nodes we operate in Tokyo, next to the validator set.
Order placement still goes through Hyperliquid's own API, so most teams run both: market data from us, orders through Hyperliquid.
How do I move an existing client over?
Streams that also exist in Hyperliquid's public API keep the same names and core payloads, so most of the change is in the request format. Subscriptions take a subscriptionId, markets and accounts go in arrays (coins, users), and the heartbeat is a WebSocket ping frame instead of a JSON message:
json
Copied
{ "method": "subscribe", "subscriptionId": "books", "subscription": { "type": "l2BookDiff", "coins": ["BTC", "ETH"], "nLevels": 20 } }
Our docs page on migrating from the Hyperliquid API lists every difference, and the connection lifecycle and replay and recovery pages cover reconnects, sequence numbers, and gaps.
Who is HyperCore streaming for?
We built it for teams whose systems depend on Hyperliquid market data:
- Trading teams and market makers that keep local books across many markets and need order-level detail
- Liquidation, arbitrage, and risk systems that follow fills and liquidations as they happen
- Wallets, brokers, and builder-code front ends that show live Hyperliquid markets to their own users and track fills routed through their builder code
- Teams already on HyperEVM with us that want HyperCore data on the same API key
If you're still comparing providers, our guide to the best Hyperliquid RPC providers covers HyperCore data and HyperEVM RPC side by side.
How do I get started?
Start with the HyperCore docs. They cover the endpoint, every stream's subscription and message format, and the recovery model. Existing Alchemy teams can use the API key they already have, and new teams can create an Alchemy account to get one. HyperCore is free for all users during launch week.
We're also building the rest of the HyperCore data stack on top of this, and the teams streaming today are shaping what comes next. If there's a stream or a feature your system needs, tell us.
