mcp -juejin

This MCP server enables AI-assisted creation and publishing of articles to the Juejin community. It exposes a tool named saveArticleToJueJin that AI agents can invoke to create a draft and publish content. Content supports Markdown formatting and tagging, and the service communicates with the Juejin API to create drafts first and then publish them when requested. When running in stdio mode, the server interacts via structured JSON messages, allowing you to pass a title, Markdown content, and tags to produce a published article (or a draft) on Juejin.

To use it, first start the MCP server in stdio mode. Then, through the MCP tooling interface, call the saveArticleToJueJin tool by providing the required arguments, such as title, markdowncontent, and tags. The tool will handle converting the supplied Markdown to the appropriate payload, call the Juejin API to create a draft, and optionally publish the article. You’ll receive a response containing the article ID and link if the operation succeeds.

Prerequisites:

• JDK 17 or newer
• Maven 3.6+ or an equivalent Java build tool
• Internet access for dependency resolution

Step 1: Download or clone the repository

• git clone https://github.com/corgiboyg/mcp-server-juejin.git
• cd mcp-server-juejin

Step 2: Build the project

• mvn clean package

Step 3: Run the MCP server (stdio mode)

• java -jar target/mcp-server-juejin-1.0.0.jar

Step 4: Verify startup

• Check console output for a successful startup message and health indicators. The server runs in stdio mode, so it will await JSON-based tool invocation requests from the MCP framework or an intelligence agent.

Notes:

• Ensure you provide a valid juejin.cookie in application.yml as described in the config example to authenticate with the Juejin API.
• If you need to adjust memory, modify JAVA_OPTS accordingly and re-run the jar.

Tips and caveats:

• Cookie validity: Juejin cookies expire; update the cookie in src/main/resources/application.yml when needed.
• Rate limits: The Juejin API may enforce rate limits; implement retry/backoff logic in your AI tooling when invoking saveArticleToJueJin.
• Markdown support: The server accepts Markdown content and uses a converter (MarkdownConverter) to prepare payloads for the Juejin API.
• Tags: Supply a comma-separated string for multiple tags; the tool will map them to the article tags on Juejin.
• stdio mode: The server is designed to run without a web container in standard input/output mode, enabling tight integration with MCP tooling.
• Logs: Review data/log/mcp-server-juejin.log for troubleshooting and operation history.