China Macro MCP Server

@pipeworx/china-macro

China's PBOC Loan Prime Rate (贷款市场报价利率, LPR), National Bureau of Statistics headline indicators, 70-city house prices, SHIBOR and the government-bond yield curve — all keyless, sourced live.

Part of Pipeworx — an MCP gateway connecting AI agents to 1863+ live data sources. This is an independent, unofficial integration — not affiliated with, endorsed by, or published by the upstream provider.

Tools

• china_lpr({ months? }) — up to the last 12 monthly LPR announcements (1-year and 5-year rate, plus reference date), newest first, and a latest field. The upstream only exposes a rolling ~1-year window — months caps at 12.
• china_lpr_latest() — the current 1-year/5-year LPR print and its reference date only, without the history. Cheaper when a caller just wants "the current rate".
• china_macro_indicator({ indicator, periods? }) — one tool across nine headline indicators: cpi, ppi, pmi (manufacturing + non-manufacturing), gdp (year-to-date cumulative, by sector), industrial_production (industrial value-added growth), retail_sales, fixed_asset_investment, money_supply (M0/M1/M2), and trade (monthly export/import totals). Returns the latest period first, then up to periods (default 12, max 36) months/quarters of history — every field translated from its Chinese NBS name to an English name with its unit. Direct route (source: "nbs"). Reads the National Bureau of Statistics of China's own database instead of the republisher. Each series comes back with its period, period_type (single_month / single_quarter / year_to_date), measure (index, pct_change_yoy, pct_change_qoq, change_pp, level, level_pct, rate_pct, diffusion_index), unit and base (e.g. "same month of the prior year = 100"), exactly as published and never converted; the publisher's own unit string is kept as published_unit because it can be wrong (a base-100 index published as "%"). A series whose base cannot be established from the publisher's label returns basis_status: "unresolved" with the number only as value_raw. Blank periods are dropped and counted, never returned as nulls. Also adds urban_unemployment and capacity_utilization (quarterly), series (one named series), period, and region: a province name for its history, or "all" for the 31 mainland provincial-level regions in one period (cpi, ppi, industrial_production, fixed_asset_investment, gdp). Taiwan, Hong Kong and Macao are not in these series.
• china_house_prices({ city?, month? }) — the 70-city new-build and resale residential price indexes. city (English name) for one city's history (up to 24 months); month ("YYYY-MM") to pin a release; neither for the latest month's full 70-city cross-section.
• china_rates({ series, days?, periods? }) — series: "shibor" for the Shanghai Interbank Offered Rate across all 8 tenors (latest fixing + daily history); series: "yield_curve" for the China Government Bond yield curve (dated 1-year/10-year benchmark history plus the full real-time quoted curve across all actively-traded maturities).

Auth

Keyless. No cookie or session.

Data sources

• National Interbank Funding Center, under PBOC authorization — the Loan Prime Rate (china_lpr, china_lpr_latest).
• National Bureau of Statistics of China — CPI, PPI, PMI, GDP, industrial value-added growth, retail sales, fixed-asset investment, money supply, and the 70-city residential price index (china_macro_indicator, china_house_prices).
• National Bureau of Statistics of China, national and provincial statistics database -- the direct route (source: "nbs", region): national monthly and quarterly series and the provincial cross-sections. data_as_of is the latest period with a value.
• General Administration of Customs of China — monthly export/import totals (china_macro_indicator({ indicator: "trade" })).
• China Foreign Exchange Trade System / National Interbank Funding Center — the SHIBOR fixing and China Government Bond yields (china_rates).

The LPR publisher exposes only a rolling ~12-month window, so china_lpr cannot return deeper history. LPR is announced monthly on the 20th (or the next business day) — expect china_lpr_latest's reference_date to move roughly once a month.

The NBS and Customs figures reach us through a keyless republisher (Eastmoney's data centre) that carries the official series; the NBS or Customs name stays first in every response's source field and the intermediary second. The 1-year/10-year government-bond benchmark is the daily series disclosed under the IMF's Special Data Dissemination Standard; it is capped at 60 rows per call here — request periods close to what you need rather than the daily maximum. The real-time yield curve defaults to Treasury bonds.

Quick Start

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

{
  "mcpServers": {
    "china-macro": {
      "url": "https://gateway.pipeworx.io/china-macro/mcp"
    }
  }
}

What this endpoint actually serves

tools/list at https://gateway.pipeworx.io/china-macro/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:

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

Both URLs reach the same gateway and the same 1863+ 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

curl -X POST https://gateway.pipeworx.io/v1/tools/china_lpr \
  -H 'Content-Type: application/json' \
  -d '{"months":12}'

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

Run it from npm

@pipeworx/mcp-china-macro is a local stdio bridge to the Pipeworx gateway: your MCP client starts it with npx, and it forwards every call to the same https://gateway.pipeworx.io/china-macro/mcp endpoint shown above. It needs network access; there is no offline mode.

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

Without a key it calls anonymously, capped at 50 calls a day. A free account raises that to 200 a day: get a key at https://pipeworx.io/signup and add "env": { "PIPEWORX_API_KEY": "<your key>" } to the server entry. The key is sent as a Bearer token, the same way a hosted connection sends it, so calls are metered and billed exactly as they are there.

It lists only this pack's tools by default. Set PIPEWORX_META_TOOLS=1 to also list ask_pipeworx and the other gateway-wide tools.

Behind an HTTPS proxy: Node's built-in fetch does not read HTTPS_PROXY by default. Newer Node releases honour it when NODE_USE_ENV_PROXY=1 is set; check your Node version's documentation.

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 China Macro data" })

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

More

• Docs and guides
• pipeworx.io

License

MIT