Weather MCP — Current Conditions and Forecasts for AI Agents

Weather MCP gives AI assistants and automation workflows fast, structured weather information for cities, ZIP codes, airport codes, and other location names. It runs as a Streamable HTTP Model Context Protocol (MCP) server on Apify and uses the public wttr.in weather service—no separate weather API key is required.

Use it to answer weather questions, prepare travel or field-work briefings, add conditions to daily summaries, evaluate outdoor scheduling, or give an agent weather context before it recommends what to wear or when to leave.

Available tools

Tool	What it returns
get-current-weather	Resolved location, current temperature, feels-like temperature, humidity, wind speed, conditions, and a short human-readable summary
get-forecast	Available daily forecast records with dates, high/low/average temperatures, conditions, sunrise, and sunset

Both tools accept ordinary location text such as Newark, NJ, London, 10001, or LHR. wttr.in resolves the query to its nearest known area, and the response includes that resolved city, region, and country when available.

Example tool calls

Current weather

{
  "name": "get-current-weather",
  "arguments": {
    "location": "Newark, NJ"
  }
}

Example response shape:

{
  "location": "Newark, New Jersey, United States of America",
  "temperature": "24°C",
  "feelsLike": "26°C",
  "humidity": "68%",
  "windSpeed": "11 km/h",
  "conditions": "Partly cloudy",
  "description": "Currently partly cloudy in Newark..."
}

Multi-day forecast

{
  "name": "get-forecast",
  "arguments": {
    "location": "London",
    "days": 3
  }
}

days defaults to 3 and is clamped to 1–7. The response cannot contain more days than wttr.in supplies for that location; the upstream JSON commonly provides three.

Connect from an MCP client

The Streamable HTTP endpoint is:

https://reverberant-equality--weather-mcp.apify.actor/mcp

Authenticate with your Apify API token:

Authorization: Bearer YOUR_APIFY_API_TOKEN
Content-Type: application/json
Accept: application/json, text/event-stream

Your MCP client should complete the normal initialize and notifications/initialized handshake before listing or calling tools. The server runs in Apify Standby mode so clients can invoke it as an online MCP endpoint rather than launching a one-off scraping job.

Pricing

Weather MCP uses Apify Pay Per Event pricing:

• tool-call: $0.005 per tool invocation
• Apify platform usage is handled according to Standby-mode billing

A forecast request is one tool call regardless of the requested number of days. Review the live Pricing tab for the authoritative current price before high-volume use.

Data source and limitations

Weather data comes from wttr.in. Values are returned in metric units: temperatures in Celsius and wind speed in kilometers per hour. Forecast availability and location resolution depend on the upstream service. Ambiguous place names may resolve to an unexpected nearby location, so use a state, region, country, postal code, or airport code when precision matters.

Weather MCP is suitable for general planning and informational use. It is not an official severe-weather warning service and should not be the sole source for aviation, marine, emergency, or safety-critical decisions. For hazardous weather, verify conditions with the relevant government meteorological authority.

If the upstream service is unavailable or returns no usable conditions, the MCP returns a structured tool error rather than inventing weather data.

Run locally

npm install
npm run build
npm test
npm run start:dev

The local MCP endpoint defaults to http://localhost:3000/mcp. On Apify, the server uses the platform-provided container port and exposes a readiness probe at /.

Privacy

The server sends the supplied location string to wttr.in to obtain weather data. Do not submit private addresses or sensitive location information unless that disclosure is appropriate for your use case.