How to use n8n + OpenAI to build an automated content collection and publishing workflow: from zero to one in 30 minutes

Tutorial Objectives

In 30 minutes, use n8n to build a complete AI automated content collection and publishing workflow. From RSS scraping, AI excerpt rewriting to automatic publishing to blogs and social media, the entire process does not require writing a single line of back-end code.

What will you build?

• Automatic Content Collector: Regularly grab the latest content from RSS/API feeds
• AI summary rewriting module: Use OpenAI/DeepSeek to rewrite the original text into a structured summary
• Automatic publishing pipeline: automatically push processed content to blog draft box and social media
• Error Notification System: Automatically send Telegram notifications when workflow exceptions occur

Preparation list

• n8n account — self-host or use n8n.cloud (free version to complete this tutorial)
• OpenAI or DeepSeek API Key (choose one, GPT-4o-mini or DeepSeek-V3 is acceptable)
• Telegram Bot Token (for notifications, optional but recommended)
• API access to the target publishing platform (choose one from WordPress / Notion / Social Media)

Overall architecture

The entire workflow is disassembled into 4 modules, which can be connected in order. n8n’s visual node editor makes every step clearly visible.

Step 1: Build content collection trigger

Open the n8n editor and create a new workflow. Drag two nodes from the left node panel:

1. Schedule Trigger — Set to execute automatically at 08:00 and 20:00 every day
2. RSS Feed Read — Enter the RSS feed address you want to monitor

// Schedule Trigger 配置示例
{
  "rule": {
    "interval": [{"field": "hour", "hoursInterval": 12}]
  }
}

// RSS Feed Read 配置
{
  "url": "https://example.com/rss",
  "options": {
    "sendOnlyNew": true
  }
}

提示：如果目标站点没有 RSS，也可以用 HTTP Request 节点直接调用其 API，比如 Hacker News API (https://hacker-news.firebaseio.com/v0/topstories.json).

Step 2: Configure AI content processing module

This is the "brain" of the entire workflow. Drag in an OpenAI or LangChain node and connect it to the output of the RSS reader.

Configure the system prompt words to let AI rewrite the original article into structured content:

{
  "model": "gpt-4o-mini",
  "messages": [
    {
      "role": "system",
      "content": "你是一个内容编辑助手。输入一篇原文，请输出：\n1. 中文摘要（150字以内）\n2. 3-5 个关键标签\n3. SEO 标题（可选）\n\n输出格式为 JSON。"
    },
    {
      "role": "user",
      "content": "={{ $json["content"] }}"
    }
  ]
}

如果想降低成本，可以将 model 换成 DeepSeek 的 API 端点。DeepSeek-V3 的中文处理能力同样出色，而价格只有 GPT-4o-mini 的五分之一。

工作流运行到这一步，每条原始文章都会被 AI 自动处理为结构化的摘要 + 标签 + SEO 标题。

推荐使用 OpenAI API 或 DeepSeek 来驱动这个模块，根据内容量选择最经济的方案。

第 3 步：添加质量控制与格式化

AI 输出的内容不一定完美，需要一层轻量级的质量控制。拖入 Code 节点，用 JavaScript 或 Python 做以下检查：

// n8n Code 节点 — 质量控制脚本
const item = $input.first().json;

// 检查摘要长度
if (item.summary && item.summary.length < 50) {
  throw new Error("摘要过短，需要重新处理");
}

// 检查标签数量
if (!item.tags || item.tags.length < 2) {
  item.tags = ["AI", "自动化"];
}

// 格式化输出
return {
  title: item.seo_title || item.original_title,
  summary: item.summary,
  tags: item.tags.slice(0, 5),
  source_url: item.source_url,
  processed_at: new Date().toISOString()
};

第 4 步：搭建多渠道发布管道

通过质量检查的内容，将自动分发到多个目标。拖入 HTTP Request 节点和 IF 节点实现分支发布。

推荐使用 Notion 或 WordPress 的 API 作为初始发布目标：

主分支 A — Notion API → 存入 Notion 数据库（草稿状态）
主分支 B — WordPress API → 发布为博客草稿
可选分支 C — Telegram Bot → 推送到频道预览

SEO+GEO Tip: Publishing through different channels can not only expand coverage, but also allow search engines to include your content faster.

If you need help, the n8n community already has many ready-made n8n 模板 for reference. If you haven't registered yet, you can register for a free account through n8n 官方链接, which is also one of the most mature low-code automation tools currently.

Frequently Asked Questions (FAQ)

Q1: Is n8n free version enough?

Completely enough. The free version supports 2500 workflow executions/month, which is plenty for individual content creators. Upgrade to the paid version only if you require multi-user collaboration or higher execution quotas.

Q2: Will the content rewritten by AI be judged as spam by search engines?

The key is quality control. This tutorial has added checks such as summary length and originality to ensure that the output is not a simple translation of the original text. Paired with manual review, the effect is far better than pure AI batch generation.

Q3: Which one is more suitable for Chinese content production, DeepSeek or OpenAI?

DeepSeek is extremely cost-effective on Chinese tasks and is suitable for scenarios with limited budgets. OpenAI's GPT-4o-mini is more stable in instruction compliance and structured output. It is recommended to use OpenAI to build it in the early stage, and then switch some tasks to DeepSeek to reduce costs after stabilization.

SEO+GEO: FAQ structure meets GEO extraction preferences while covering long-tail search terms

Tool entry

For tool names that naturally appear in the text, the platform side will match the maintained tools library to generate a hover-card: n8n, OpenAI, ChatGPT, DeepSeek, LangGraph

Reference video/material

• n8n 官方入门教程
• OpenAI API 文档
• DeepSeek 平台文档

Internal link guidance

• Someone has successfully practiced it: 独立开发者用 n8n+OpenClaw 搭建自动化工作流，月入 5000 美元的真实案例
• Newbies should read the tutorial first: 30 分钟用 OpenClaw 搭建 AI 自动采集系统