Travel

You are building and using a Python MCP server suite that orchestrates travel planning tasks. It exposes modular servers to read local place data, plan routes, query weather, generate travel images, and publish content to Xiaohongshu, enabling end-to-end travel planning workflows from a single MCP ecosystem.

How to use

You interact with the five MCP servers through a client that understands the MCP protocol. The servers are designed to work together so you can query places, plan routes, fetch weather, generate a vertical travel poster, and publish Xiaohongshu notes all in sequence or in batch. Each server runs as a local stdio process and communicates over the MCP channels defined below.

Typical usage patterns include: 1) look up places by province or city, 2) plan a multi-point route between attractions, 3) fetch current and forecasted weather for travel days, 4) generate a long travel guide poster image, and 5) generate Xiaohongshu content and publish it in one or more steps. You can combine these steps to create a complete travel planning workflow.

How to install

Prerequisites: install Python 3.8+ on your system and obtain API keys for the services used by the MCP stack. You also need a way to run Python commands from the command line.

Install required Python packages using a single command.

pip install mcp fastmcp httpx pydantic selenium

Server configuration and startup steps

Next, create configuration entries that point to the local Python scripts for each MCP server. The following examples show the explicit commands and paths you should use. Adjust the placeholders to your actual Python interpreter path and project root.

{
  "mcpServers": {
    "amap_route_planning": {
      "autoApprove": ["multi_point_route", "route_planning", "search_places"],
      "disabled": false,
      "timeout": 600,
      "type": "stdio",
      "command": "python",
      "args": ["你的项目路径\\middleware\\route_planning_mcp.py"],
      "cwd": "你的项目路径",
      "env": {
        "AMAP_API_KEY": "YOUR_AMAP_API_KEY"
      }
    },
    "amap_weather": {
      "autoApprove": ["get_current_weather", "get_weather_forecast", "get_complete_weather", "search_city_weather"],
      "disabled": false,
      "timeout": 300,
      "type": "stdio",
      "command": "python",
      "args": ["你的项目路径\\crawler\\weather_mcp.py"],
      "cwd": "你的项目路径",
      "env": {
        "AMAP_API_KEY": "YOUR_AMAP_API_KEY"
      }
    },
    "places_read": {
      "autoApprove": ["get_spots_by_province", "get_spots_by_city", "get_all_provinces", "get_cities_in_province", "search_spots_by_keyword"],
      "disabled": false,
      "timeout": 300,
      "type": "stdio",
      "command": "python",
      "args": ["你的项目路径\\crawler\\places_read_mcp.py"],
      "cwd": "你的项目路径"
    },
    "image_generator": {
      "autoApprove": ["generate_image_nano_banana"],
      "disabled": false,
      "timeout": 300,
      "type": "stdio",
      "command": "python",
      "args": ["你的项目路径\\middleware\\generate_mcp.py"],
      "cwd": "你的项目路径"
    },
    "xiaohongshu_publisher": {
      "autoApprove": ["publish_xiaohongshu_video", "publish_xiaohongshu_images", "generate_xiaohongshu_content", "batch_publish_xiaohongshu"],
      "disabled": false,
      "timeout": 300,
      "type": "stdio",
      "command": "python",
      "args": ["你的项目路径\\publisher\\publish_mcp.py"],
      "cwd": "你的项目路径"
    }
  }
}

Additional configuration and notes

Environment variables shown in the configuration are required to run the servers. You should replace the placeholders with your actual keys before starting the MCP stack.

Security and operational notes: keep API keys out of version control. Use environment variables or separate configuration files to manage sensitive data. The Xiaohongshu publisher server requires a logged-in browser session and corresponding ChromeDriver setup when publishing content.

Troubleshooting and tips

If a server fails to start, verify that the Python interpreter path and project root in the configuration are correct, ensure all dependencies are installed, and check the server logs for detailed error messages.

If API calls fail, confirm that your API keys are valid and properly exposed to the environment variables used by the servers.

Project structure overview

Core components are organized to separate concerns: data reading, weather, route planning, image generation, and Xiaohongshu publishing. Each module exposes a focused set of tools that you can call from your MCP client to compose end-to-end travel workflows.

Available tools

get_spots_by_province

Retrieve all spots within a specified province.

get_spots_by_city

Retrieve all spots within a specified city.

get_all_provinces

List all provinces with available spot data.

get_cities_in_province

List all cities in a given province.

search_spots_by_keyword

Search spots by a keyword term.

travel_image_prompt_guide

Generate a prompt framework for travel posters.

generate_image_nano_banana

Call Nano Banana API to create a travel poster image.

publish_xiaohongshu_video

Publish a travel video note to Xiaohongshu.

publish_xiaohongshu_images

Publish a travel image note to Xiaohongshu.

generate_xiaohongshu_content

Generate Xiaohongshu content (title, body, topics) from location data.

batch_publish_xiaohongshu

Publish multiple Xiaohongshu notes in batch.

route_planning

Plan a route for a chosen travel mode.

search_places

Search for places to visit.

multi_point_route

Compute routes that visit multiple points.

get_current_weather

Query real-time weather for a location.

get_weather_forecast

Query weather forecast for a location.

get_complete_weather

Query full weather data (real-time + forecast).

search_city_weather

Search for a city and retrieve its weather.