Overview
What is blogger-mcp-server?
A Model Context Protocol (MCP) server that allows AI models like Claude to interact with Google's Blogger API. It enables listing and retrieving blogs, managing posts (CRUD), and listing labels. It uses API key authentication but cannot create new blogs (Google limitation).
How to use blogger-mcp-server?
Install via npm (npm install -g @mcproadev/blogger-mcp-server) or from source. Configure the environment variable BLOGGER_API_KEY with your Google API key. Use with any MCP client (e.g., Claude Desktop) by providing the server configuration in your client's settings. Start with ./start-dev.sh (development) or ./start-prod.sh (production).
Key features of blogger-mcp-server
- List and retrieve blogs
- List, search, retrieve, create, update, delete posts
- List and retrieve labels
- API key authentication only
- Deployable to Vercel, Docker, Heroku, AWS Lambda, Google Cloud Run
Use cases of blogger-mcp-server
- AI assistant managing blog content via natural language commands
- Automating blog post creation, updates, and deletions
- Searching and filtering blog posts for specific keywords
- Integrating Blogger with other MCP-compatible tools and platforms
FAQ from blogger-mcp-server
Can I create a new blog via this server?
No. Google's Blogger API does not allow creating blogs programmatically. Blogs must be created manually through the Blogger web interface.
What are the prerequisites?
Node.js version 16 or higher and a Google Blogger API key obtained from the Google Cloud Console.
How is authentication handled?
The server uses API key authentication only. Set the environment variable BLOGGER_API_KEY with your key. This limits access to public blogs or blogs explicitly configured for API access.
Can I search posts?
Yes, but the Blogger API lacks a direct search endpoint. The server implements search client-side by fetching all posts and then filtering them based on your query.
What deployment options are available?
The server can be deployed on Vercel (using vercel.json), Docker (using the included Dockerfile), Heroku, AWS Lambda, or Google Cloud Run.
