{"id":37643,"date":"2025-06-22T09:24:01","date_gmt":"2025-06-22T01:24:01","guid":{"rendered":"https:\/\/www.1ai.net\/?p=37643"},"modified":"2025-06-17T12:28:23","modified_gmt":"2025-06-17T04:28:23","slug":"%e6%89%a3%e5%ad%90coze%e5%ae%9e%e6%88%98mcp%e6%95%99%e7%a8%8b%ef%bc%8c%e4%b8%80%e9%94%ae%e7%94%9f%e6%88%90%e5%85%ac%e4%bc%97%e5%8f%b7%e6%96%87%e7%ab%a0%e3%80%81%e8%87%aa%e5%8a%a8%e6%8e%92%e7%89%88","status":"publish","type":"post","link":"https:\/\/www.1ai.net\/en\/37643.html","title":{"rendered":"Buckle Coze real-world MCP tutorial, one-click generation of public articles, automatic layout, automatic publishing"},"content":{"rendered":"

In the last few days I've seen an automated generationPublic Account<\/a>writingWorkflow<\/a>The feeling is very practical value, hurry up to introduce to you, they are also as notes to keep, good memory as bad pen ah, haha, not much to say, on the dry goods.<\/p>\n

Overall Introduction<\/strong><\/p>\n

This mcp service, divided into three parts, consists of three workflows:<\/p>\n

I. Article generation workflow<\/p>\n

II. Article layout workflow<\/p>\n

Third, the article is automatically published to the public workflow<\/p>\n

Workflow Details<\/strong><\/p>\n

I. Article generation workflow<\/p>\n

\"719e5a93j00sxzg3i001gd000u000acm\"<\/p>\n

Article generation workflow is actually relatively simple, is the use of large models of the ability to generate, given a topic, and Prompt settings, large models to generate, specific Prompt is as follows:<\/p>\n

You are a master at writing an in-depth article based on a topic provided by a user that<\/p>\n

and the typography is perfect. the article is microsoft style.<\/p>\n

This is something you can optimize on your own, or you can add some crawlers and online networking features to get the information and then write the article to add some real corpus to the article<\/p>\n

II. Article layout workflow<\/p>\n

Article layout is relatively complex a little bit, the core is to generate html article content format, public support for some html tags<\/p>\n

\"d8772d81j00sxzg3i0026d000u000bvm\"<\/p>\n

First input node, input the output content of the previous workflow, the articulation of this through the back of the release of the mcp service to do, do not worry about it here first.<\/p>\n

The large model part of the use of deepseek-v3-0324 model, this model generates a good effect, the length of the article is long enough, there is a point of attention here, you need to set the length of the large model generation<\/p>\n

\"6eb4c556j00sxzg3i001zd000j800hem\"<\/p>\n

Other settings are defaulted and you can explore them as well<\/p>\n

The large model section also requires a cue word section, as follows<\/p>\n

\"e8e9af63j00sxzg3h0021d0008u0096m\"<\/p>\n

System prompts and user prompts, just follow my design!<\/p>\n

After the big model to add a code node, mainly to return the dictionary format text, this simple<\/p>\n

Third, the article is automatically published to the public workflow<\/p>\n

And finally the workflow for posting to the public<\/p>\n

\"ad453d28j00sxzg3i003dd000u0009gm\"<\/p>\n

First of all, the nodes needed for public release require three plug-ins, and the roles of these three plug-ins are described below:<\/p>\n