Yandex Weather MCP
A local MCP server for the Yandex Weather API: current conditions, forecast, a short summary and comparing two locations, with caching
Low risk
We rate an entry low when it mostly gives the agent instructions and reference material.
Why this level
- Only reads public weather data and does not modify anything in external systems
Install
Manual install
npm install && npm run build && npm startLocal run after cp .env.example .env and filling in YANDEX_WEATHER_API_KEY.
This is third-party code. Review the repository files before installing.
What it does
The server hides the Yandex Weather key from the MCP client and provides four tools: current weather, a multi-day forecast, a short text summary with advice, and a comparison of two locations. API responses are normalized and cached for 5 to 15 minutes, and input coordinates are validated with Zod schemas. It can run locally via npm or in Docker with HTTP transport on /mcp and a health check, with a sample Nginx config for VDS deployment.
Who it is for. For people who want an agent to check the weather by coordinates instead of opening a weather app
Good fit when
- You want an agent to answer current weather or forecast questions by coordinates
- You need to compare weather in two cities in one request
- You want a self-hosted weather MCP server instead of a third-party cloud one
Not a fit when
- You need a ready public server without your own key and deployment
- You need hourly data not available on your Yandex Weather plan
Example request
Compare the weather in Moscow and Saint Petersburg right now and tell me where it is warmerLimitations
The project has no stars and, per the README, was written by a single author using a role-play write-up style to explain decisions. It needs your own Yandex Weather API key and your own hosting or a local run; there is no ready public endpoint. There is no stable release or npm version.
How to disable. Stop the container or server process and remove it from your MCP client config.
MCP
- Transport
- stdio, http
- Authentication
- API key
| Environment variables | |
|---|---|
| YANDEX_WEATHER_API_KEY required, secret | Yandex Weather API key |
| YANDEX_WEATHER_LANG | Response language, e.g. ru_RU |
| CACHE_TTL_SECONDS | Response cache TTL in seconds |
| MCP_TRANSPORT | MCP transport: http or stdio |
| PORT | HTTP server port, default 3000 |
Security check
- Only reads public weather data and does not modify anything in external systems
README in short
The README frames the server as the result of several role perspectives (backend, devops, integrations, MCP, quality) and explains choosing TypeScript and Node.js. It covers .env setup, a local run, Docker with HTTP transport on /mcp, four MCP tools with JSON examples, and a VDS deployment guide with Nginx and firewall notes.
FAQ
Where do I get an API key?
In the Yandex Weather or Yandex API personal cabinet; the endpoint is set via YANDEX_WEATHER_API_URL if your plan uses a different address.
Can I run it without Docker?
Yes, with npm install, npm run build and npm start, setting MCP_TRANSPORT=stdio for a local MCP client.
Related
A self-hosted knowledge base with block-level references and a built-in MCP server for connecting AI agents to your notes
A CLI for every Google Workspace API with JSON output and agent skills: Drive, Gmail, Calendar, Sheets and more
Local search over Markdown notes, docs and meeting transcripts: keywords, semantic search and reranking, with an MCP server
A task manager for AI-driven development: breaks a PRD into dependent tasks and guides the agent through them via MCP or CLI