Hyperliquid

io.github.pipeworx-iov0.1.0更新於 Oct 3, 2026

Hyperliquid MCP — read-only market data from Hyperliquid's public info API.

已驗證Streamable HTTP可網頁執行FinanceData & Analytics

概覽

AI 產生的概覽

為助理提供唯讀的 Hyperliquid 市場資料:中間價、訂單簿、K 線、資金費率與未平倉量。

功能
把 Hyperliquid 的公開 info API 包裝成唯讀工具:永續合約中繼資料、現貨交易對中繼資料、全部符號的中間價、單一資產脈絡(資金費率、未平倉量、標記/預言機/中間價、溢價、24 小時成交量)、L2 訂單簿、OHLCV K 線,以及已實現資金費率歷史。資料依請求即時代理,刻意不包含錢包、簽章與交易端點。閘道端點還會列出 ask_pipeworx、discover_tools 等共用中繼工具。
適用情境
當助理需要 Hyperliquid 永續與現貨的即時或歷史行情資料,用於分析、監控或報告,且不需要交易或帳戶存取時使用。不適合下單或讀取私有帳戶狀態。
執行需求
使用 gateway.pipeworx.io 上的遠端 streamable HTTP 端點;讀取不需要帳號、API 金鑰或簽章。也可透過 npx @pipeworx/mcp-hyperliquid 以本機 stdio 方式執行,需要 Node.js。需要能連線到閘道與 Hyperliquid 公開 info 端點。
安裝前請注意
設計上為唯讀:不涉及錢包、簽章或交易。資料來自第三方閘道與 Hyperliquid 公開 API,可用性與速率限制不受你控制;觸發限流時會回傳上游原始錯誤。閘道端點還會暴露共用中繼工具,會增加載入脈絡的工具結構描述數量。這是非官方整合,與 Hyperliquid 無隸屬關係。

安裝

在 SourceWeft 中

  1. 開啟 儀表板中的 Hyperliquid,將其新增到工作區。
  2. 為需要使用其工具的對話啟用該服務。

Web executable,透過 Streamable HTTP。 遠端服務在工作區中設定後即可從網頁執行環境執行。

其他 MCP 客戶端

把它新增到你客戶端的 mcpServers 設定中。

{
  "mcpServers": {
    "hyperliquid": {
      "type": "http",
      "url": "https://gateway.pipeworx.io/hyperliquid/mcp"
    }
  }
}

README

@pipeworx/hyperliquid

Read-only market data for Hyperliquid, the largest onchain perpetuals exchange (it also runs a spot market) — mid prices, order books, candles, funding rates/history and live open interest — proxied live, per request, from Hyperliquid's own public info API. No wallet, no signing, no trading: the exchange/signed-order endpoints are out of scope on purpose.

Part of Pipeworx — an MCP gateway connecting AI agents to 1688+ live data sources. This is an independent, unofficial integration — not affiliated with, endorsed by, or published by the upstream provider.

Tools

  • hyperliquid_perp_meta(coin?) — the perps universe (symbol, size decimals, max leverage, delisted flag). Pass coin for one symbol, omit for the whole list.
  • hyperliquid_spot_meta() — every spot trading pair Hyperliquid lists, each with its pair name, pair index, and the token ids it is composed of.
  • hyperliquid_mids(coin?) — current mid price for every perp/spot symbol in one call, or just one symbol's mid price.
  • hyperliquid_asset_context(coin) — a perp's live funding rate, open interest, mark/oracle/mid price, premium, and 24h volume.
  • hyperliquid_l2_book(coin, depth?) — the live bid/ask order book (price, aggregate size, order count per level).
  • hyperliquid_candles(coin, interval?, lookback_hours?, start_time?, end_time?) — OHLCV candles for any window Hyperliquid retains.
  • hyperliquid_funding_history(coin, lookback_hours?, start_time?, end_time?) — the realized funding-rate series (Hyperliquid pays funding hourly).

Auth

Keyless. Hyperliquid's info endpoint (POST api.hyperliquid.xyz/info) takes no API key, account, or signature for any read in this pack.

Data sources

  • https://api.hyperliquid.xyz/info — Hyperliquid's public info endpoint. Every tool above sends a different type in the POST body (meta, spotMeta, allMids, metaAndAssetCtxs, l2Book, candleSnapshot, fundingHistory) — there is one URL, not one per resource.

Things the next caller would otherwise rediscover the hard way:

  • No per-request rate-limit error observed in testing — Hyperliquid's public docs describe IP-based weight limits on the info endpoint, but this pack does not special-case a 429; a caller hitting one gets Hyperliquid's own error body verbatim (via httpError), not a Pipeworx-specific message.
  • Perp vs spot coin IDs are DIFFERENT NAMESPACES, same coin argument shape. A perp is addressed by its bare symbol ("BTC", "ETH"). A spot pair is addressed by "@<index>" (e.g. "@0") EXCEPT the one canonical pair per token, which also has a human name ("PURR/USDC"). Get the right id from hyperliquid_spot_meta's index field — guessing a spot pair's @N id from its position elsewhere will be wrong, since indices are not contiguous by listing date.
  • allMids keys are a flat mix of perp symbols and spot @N ids — there is no type field distinguishing them in the response; you have to already know which namespace a symbol belongs to (perp list from hyperliquid_perp_meta, spot list from hyperliquid_spot_meta).
  • candleSnapshot returns [] (not an error) for a startTime/endTime window that contains no interval boundary — e.g. a 100ms window at 1h granularity. This is a legitimate empty answer, not a broken call; widen the window or use a coarser interval.
  • Open interest is reported in BASE ASSET UNITS, not USD (e.g. a BTC figure like "37751.7" means 37,751.7 BTC of open interest, not dollars) — multiply by markPx for a USD figure if that's what's needed.
  • metaAndAssetCtxs returns a two-element array, [meta, assetCtxs], where assetCtxs[i] corresponds to meta.universe[i] — there is no per-asset key; you must look up the coin's index in universe first and use that same index into assetCtxs. This pack does that lookup for you in hyperliquid_asset_context.
  • Historical open-interest is not published separately — only the CURRENT open interest is available (via metaAndAssetCtxs/ hyperliquid_asset_context); there is no openInterestHistory endpoint analogous to fundingHistory.

Quick Start

Add to your MCP client (Claude Desktop, Cursor, Windsurf, etc.):

json
{  "mcpServers": {    "hyperliquid": {      "url": "https://gateway.pipeworx.io/hyperliquid/mcp"    }  }}

What this endpoint actually serves

tools/list at https://gateway.pipeworx.io/hyperliquid/mcp returns the tools in the table above plus the shared Pipeworx meta-tools — ask_pipeworx, discover_tools, search_within, remember/recall and the rest of the gateway-wide set. So the tool count you see is larger than this table: a single-pack endpoint currently lists roughly 30 shared tools alongside the pack's own. The connection's initialize response states its exact scope, and is the authoritative answer for a given day.

This is deliberate, not multiplexing by accident. The meta-tools are what let a scoped connection answer a question this pack does not cover — via ask_pipeworx, which routes across the whole catalog — without you adding a second MCP server. There is currently no way to mount a pack endpoint without them; if the extra schemas cost you more context than the routing is worth, connect to the full gateway once rather than to several pack endpoints.

Or connect to the full Pipeworx gateway to get every pack's tools listed directly, instead of just this one's:

json
{  "mcpServers": {    "pipeworx": {      "url": "https://gateway.pipeworx.io/mcp"    }  }}

Both URLs reach the same gateway and the same 1688+ data sources. The only difference is which pack's tools are listed directly; ask_pipeworx reaches all of them from either one.

No MCP client? Call it over HTTP

bash
curl -X POST https://gateway.pipeworx.io/v1/tools/hyperliquid_perp_meta \  -H 'Content-Type: application/json' \  -d '{"coin":"BTC"}'

No account needed for the first calls. Inspect any tool: GET https://gateway.pipeworx.io/v1/tools/hyperliquid_perp_meta. Find one: POST https://gateway.pipeworx.io/v1/tools/search_packs with {"query":"..."}.

Standalone (no gateway account)

This package also runs as a local stdio MCP server — no Pipeworx account, no gateway round-trip:

json
{  "mcpServers": {    "hyperliquid": {      "command": "npx",      "args": ["-y", "@pipeworx/mcp-hyperliquid"]    }  }}

Or run it directly to confirm it starts:

bash
npx -y @pipeworx/mcp-hyperliquid

It speaks MCP over stdin/stdout and answers initialize/tools/list/tools/call for only this pack's tools — none of the shared meta-tools the gateway connection above adds. Same source, same tools, no ask_pipeworx routing.

Using with ask_pipeworx

Instead of calling tools directly, you can ask questions in plain English — this works on the pack endpoint above as well as on the full gateway:

ask_pipeworx({ question: "your question about Hyperliquid data" })

The gateway picks the right tool and fills the arguments automatically.

More

License

MIT

來源:README.md,提交 fcfd094

工具

0
工具後設資料尚未被收錄。

版本歷史

1
  1. v0.1.0最新Oct 3, 2026