MCP服务器从5万条推文和400幅视觉艺术作品中生成文章。一个命令。三个文章概念。选一个数字。

为什么

无论我们是否参与，人工智能都在写关于我们想法的文章。因此，我们将源材料——5万条推文、400条视觉图和精确的写作档案——构建成了一个任何人都能插入AI代理的工具。

安装

克劳德·科德

克劳德桌面 / 光标 / 风帆冲浪

添加你的MCP配置：

它是如何运作的

1. 无参数叫vvriter
2. 从3个文章概念中选择
3. 文章保存到 ~/vvriter/ 并在浏览器中打开

该工具加载随机样本的推文和视觉图像。AI会找到想法集群。每次通话都会随机排列——你永远不会收到相同的建议两次。

它是如何建造的

完整技术解析：如何构建一个用于AI驱动内容生成的MCP服务器

相关

• jackbutcher.md — 驱动声音的写作配置文件
• 写作画像的制作过程

关于可视化价值

杰克·布彻的艺术、课程和工具。5门课程，168节课，55,000+名学生，607条五星好评。

• 课程——杠杆、价值创造、独立建设
• 工作流程——人工智能如何构建事物

vvriter

MCP server that generates articles from 50,000 tweets and 400 visual artworks. One command. Three article concepts. Pick a number.

Why

AI is writing articles about our ideas whether we participate or not. So we built the source material — 50,000 tweets, 400 visuals, and an exact writing profile — into a tool anyone can plug into their AI agent.

Install

npx vvriter

Claude Code

claude mcp add vvriter -- npx vvriter

Claude Desktop / Cursor / Windsurf

Add to your MCP config:

{
  "mcpServers": {
    "vvriter": {
      "command": "npx",
      "args": ["vvriter"]
    }
  }
}

How it works

1. Call vvriter with no arguments
2. Pick from 3 article concepts
3. Article saves to ~/vvriter/ and opens in your browser

The tool loads a randomized sample of tweets and visuals. The AI finds idea clusters. Every call shuffles — you never get the same suggestions twice.

How it was built

Full technical breakdown: How to Build an MCP Server for AI-Powered Content Generation

Related

• jackbutcher.md — the writing profile that powers the voice
• How the writing profile was made

About Visualize Value

Art, courses, and tools by Jack Butcher. 5 courses, 168 lessons, 55,000+ students, 607 five-star reviews.

• Courses — leverage, value creation, building independently
• Workflows — how things get built with AI
• Art — Checks ($250M+), Opepen, Self Checkout
• Visuals — 600+ visual artworks
• The $99 MBA — all 5 courses for $99/year

License

MIT