feat: mcp server, ai guide (#36)

* feat: mcp server, ai guide

* feat: add context7 json file

Author: clampr

docs/ai.md

@@ -0,0 +1,19 @@
+---
+sidebar_position: 9
+---
+
+# Artificial Intelligence
+
+Meteostat provides different interfaces for accessing information using generative AI. Here you can find more information about these interfaces and how to use them.
+
+## Skill
+
+Meteostat provides a [free skill](/cli/skill) based on the [Meteostat CLI](/cli). This allows you to ask questions about weather and climate data in natural language. It is primarily designed for local use with AI agents, such as Claude, GitHub Copilot, OpenCode, Kiro, etc.
+
+## MCP Server
+
+If you're looking for a more scalable solution, Meteostat also provides a [remote MCP server](/api/mcp) through RapidAPI. This provides access to weather and climate data through AI agents that support the MCP protocol, enabling you to ask questions about historical weather data in natural language.
+
+## Context7
+
+If you want to enable your AI agent to access the entire Meteostat documentation, use [Context7](https://context7.com/websites/dev_meteostat_net). You'll need to [setup Context7](https://github.com/upstash/context7#installation) with your AI agent. Then, you can ask questions about the documentation in natural language. This is especially useful if you're building software that uses one of the Meteostat interfaces and you want to quickly find relevant information about the API endpoints, parameters, etc.

docs/api/mcp.md

@@ -0,0 +1,41 @@
+---
+title: MCP Server | JSON API
+sidebar_label: MCP Server
+sidebar_position: 4
+---
+
+# MCP Server
+
+Meteostat provides a remote MCP (Model Context Protocol) server through RapidAPI. This provides access to weather and climate data through AI agents that support the MCP protocol, enabling you to ask questions about historical weather data in natural language.
+
+## 📚 Installation {#installation}
+
+Before you can start asking questions about the weather, you need to [set up an account on RapidAPI and subscribe to the Meteostat API](https://rapidapi.com/meteostat/api/meteostat). Then, you must add the MCP server to your agent's configuration. The exact steps depend on the agent (Claude, GitHub Copilot, OpenCode, Kiro, etc.) you are using, but generally involve adding something like the following to your agent's configuration file:
+
+```json
+{
+  "mcpServers": {
+    "meteostat": {
+      "command": "npx",
+      "args": [
+        "mcp-remote",
+        "https://mcp.rapidapi.com",
+        "--header",
+        "x-api-host: meteostat.p.rapidapi.com",
+        "--header",
+        "x-api-key: YOUR_RAPIDAPI_KEY"
+      ]
+    }
+  }
+}
+```
+
+## 🚀 Usage {#usage}
+
+Now you can ask questions about weather and climate data in natural language. For example, you can ask:
+
+> What was the hottest day in Frankfurt in the 2020s?
+
+The AI agent will then use the MCP server to query the Meteostat API and provide you with an answer. Something like this:
+
+> The hottest day in Frankfurt in the 2020s (through 2025) was August 4, 2022, with a maximum temperature of 38.6°C (101.5°F). That was part of a severe European heat wave that summer.

docs/faq.md

@@ -1,5 +1,5 @@
 ---
-sidebar_position: 9
+sidebar_position: 10
 ---
 
 # FAQ

sidebars.ts

@@ -55,6 +55,10 @@ const sidebars: SidebarsConfig = {
       type: "doc",
       id: "donate",
     },
+    {
+      type: "doc",
+      id: "ai",
+    },
     {
       type: "doc",
       id: "faq",

static/context7.json

@@ -0,0 +1,4 @@
+{
+  "url": "https://context7.com/websites/dev_meteostat_net",
+  "public_key": "pk_C7CObPQVIFduy2xqz2jUH"
+}