Trendy Post MCP 🚀✨

A Model Context Protocol (MCP) service that:

1. Takes a screenshot from the user 📸

2. Uses OCR to extract text and analyze the image 🔍

3. Generates a trending Xiaohongshu-style post based on the content 📱💬

Features ✨

• Image Processing 🖼️: Extract text and analyze images

• Content Generation 📝: Create engaging Xiaohongshu-style posts

• LLM Integration 🧠: Uses AI to generate post styles, hashtags, and titles

• MCP Compatibility 🔌: Works with any MCP client

Related MCP server: RedNote MCP

Setup 🛠️

This MCP uses the trendy_post_mcp conda environment with Python 3.12 and Surya OCR.

# Activate the conda environment
conda activate trendy_post_mcp

# Run the MCP server
python server.py

Usage 📋

Once the server is running, it can be used as an MCP service that provides functions for:

• Processing screenshots 📸

• Analyzing image content 🔍

• Generating trending social media posts in the style of Xiaohongshu 📱

MCP Functions

The server exposes the following MCP functions:

• process_screenshot: Extract text from an image URL 🔤

• generate_post: Create a Xiaohongshu post from image analysis data ✍️

• process_and_generate: Combine both functions in one step (recommended) 🔄

• health_check: Check if the server is running properly 🩺

Project Structure 📁

• server.py: Main MCP server implementation

• image_processor.py: Screenshot processing and OCR functionality

• post_generator.py: Xiaohongshu post generation logic

• requirements.txt: Python dependencies

Dependencies 📦

OCR Engine

This project uses Surya OCR for text extraction, which is licensed under the GPL-3.0 License. Surya is a powerful OCR engine that supports multiple languages and provides excellent results for complex layouts.

⚠️ Important License Note: As Surya is licensed under GPL-3.0, any distribution of this software must comply with the GPL-3.0 license terms. Please ensure you understand these terms if you plan to distribute or modify this software.

Other Dependencies

• FastMCP: For MCP server implementation

• Pydantic: For data validation

• Pillow: For image processing

• ZhipuAI: For LLM-based content generation

License 📄

This project is licensed under the MIT License - see the LICENSE file for details. (Note: Components like Surya OCR have their own licenses as mentioned above)