China Air Quality MCP Server

@pipeworx/china-air-quality

Real-time nationwide air quality (空气质量) for China — AQI, pollutant levels and concentrations (PM2.5, PM10, O3, NO2, SO2, CO), quality rating and monitoring-station detail. Keyless.

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.

Of the 20 China-government sources probed 2026-09-07 (docs/china-vertical-plan.md §2), this was the only one that answered clean JSON with no session cookie, Referer trick or undocumented catalog id. No external demand asked for it in the prior 30 days; it is built because it is distinctive (nobody else serves CNEMC over MCP), not because a question was waiting on it.

Tools

• china_air_quality(city) — a city's current AQI, quality rating, primary pollutant, per-pollutant index level, and averaged station concentrations. Accepts an English city name ("Beijing"), a province ("Guangdong" resolves to its capital, Guangzhou), or the Chinese name directly ("北京市").
• china_air_quality_ranking(direction?, n?, pollutant?) — nationwide ranking of all 338 CNEMC-monitored cities by AQI or by one pollutant's index level. direction: "bottom" (default) surfaces the most polluted; "top" the cleanest.
• china_air_quality_stations(city) — every individual monitoring station within one city, with its own AQI, concentrations and coordinates — the neighborhood-level view the city-wide summary averages over.

Source

China National Environmental Monitoring Centre (中国环境监测总站), the government body that operates China's official air-quality monitoring network:

• https://air.cnemc.cn:18007/CityData/GetAllCityRealTimeAQIModels — all 338 cities, refreshed hourly. Carries AQI and a 1-6 index level per pollutant, but no raw concentration.
• https://air.cnemc.cn:18007/CityData/GetAQIDataPublishLive?cityName=<市名> — station-level rows for one city, with actual pollutant concentrations (µg/m³, CO in mg/m³). GetAQIDataPublishLives (plural) is a dead path — it 404s.

china_air_quality merges both calls: city-level AQI/level/quality from the cached all-cities feed, and averaged concentrations computed from that city's stations, because the all-cities endpoint alone never carries a concentration value.

The all-cities payload (144 KB, hourly) is cached in-isolate for 10 minutes.

City names — Chinese only on the upstream

CNEMC identifies every city by its Chinese administrative name; "Beijing" resolves nowhere on the wire. src/index.ts bundles a generated table:

• CITY_EN_TO_ZH — Hanyu Pinyin romanization of all 338 published city names (administrative suffix 市/州/盟/地区/自治州 stripped), with a hand override for the handful whose common English spelling diverges from straight pinyin: Xi'an, Harbin, Urumqi, Lhasa, Hohhot, Ordos, Qiqihar, Hulunbuir.
• PROVINCE_TO_ZH — the 31 provinces/autonomous regions/municipalities (English and Chinese name) mapped to their capital city, so "air quality in Guangdong" resolves to Guangzhou.
• Chinese input is always accepted as-is against the 338 published names, with or without the trailing 市/州.

An unresolvable city name returns found: false with a hint rather than an upstream error — CNEMC's own unknown-city response is a silent empty array, which would otherwise read as "no pollution data" instead of "city not recognized".

Known limits

• Concentrations are a station average, not a single official number. CNEMC publishes concentrations only at station grain; the city-level summary is this pack's average across that city's reporting stations for the current hour, stated in pollutant_note on every response.
• A concentration below the sensor floor is excluded, not guessed. CNEMC reports some readings as <3 (below detection limit); those are dropped from the average rather than parsed as a wrong number.
• MEE monthly city reports are HTML and out of scope for this pack — see docs/china-vertical-plan.md §2 (filed separately as a research task).

Auth

None.

Quick Start

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

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

What this endpoint actually serves

tools/list at https://gateway.pipeworx.io/china-air-quality/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_air_quality \
  -H 'Content-Type: application/json' \
  -d '{"city":"Beijing"}'

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

Run it from npm

@pipeworx/mcp-china-air-quality 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-air-quality/mcp endpoint shown above. It needs network access; there is no offline mode.

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

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 Air Quality data" })

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

More

• Docs and guides
• pipeworx.io

License

MIT