Dynamically generate HTML page from user request using OpenAI Structured Output
工作流概述
这是一个包含7个节点的中等工作流,主要用于自动化处理各种任务。
工作流源代码
{
"id": "eXiaTDyKfXpMeyLh",
"meta": {
"instanceId": "f4f5d195bb2162a0972f737368404b18be694648d365d6c6771d7b4909d28167",
"templateCredsSetupCompleted": true
},
"name": "Dynamically generate HTML page from user request using OpenAI Structured Output",
"tags": [],
"nodes": [
{
"id": "b1d9659f-4cd0-4f87-844d-32b2af1dcf13",
"name": "Respond to Webhook",
"type": "n8n-nodes-base.respondToWebhook",
"position": [
2160,
380
],
"parameters": {
"options": {
"responseHeaders": {
"entries": [
{
"name": "Content-Type",
"value": "text/html; charset=UTF-8"
}
]
}
},
"respondWith": "text",
"responseBody": "={{ $json.html }}"
},
"typeVersion": 1.1
},
{
"id": "5ca8ad3e-7702-4f07-af24-d38e94fdc4ec",
"name": "Open AI - Using Structured Output",
"type": "n8n-nodes-base.httpRequest",
"position": [
1240,
380
],
"parameters": {
"url": "https://api.openai.com/v1/chat/completions",
"method": "POST",
"options": {},
"jsonBody": "={
\"model\": \"gpt-4o-2024-08-06\",
\"messages\": [
{
\"role\": \"system\",
\"content\": \"You are a user interface designer and copy writter. Your job is to help users visualize their website ideas. You design elegant and simple webs, with professional text. You use Tailwind framework\"
},
{
\"role\": \"user\",
\"content\": \"{{ $json.query.query }}\"
}
],
\"response_format\":
{
\"type\": \"json_schema\",
\"json_schema\": {
\"name\": \"ui\",
\"description\": \"Dynamically generated UI\",
\"strict\": true,
\"schema\": {
\"type\": \"object\",
\"properties\": {
\"type\": {
\"type\": \"string\",
\"description\": \"The type of the UI component\",
\"enum\": [
\"div\",
\"span\",
\"a\",
\"p\",
\"h1\",
\"h2\",
\"h3\",
\"h4\",
\"h5\",
\"h6\",
\"ul\",
\"ol\",
\"li\",
\"img\",
\"button\",
\"input\",
\"textarea\",
\"select\",
\"option\",
\"label\",
\"form\",
\"table\",
\"thead\",
\"tbody\",
\"tr\",
\"th\",
\"td\",
\"nav\",
\"header\",
\"footer\",
\"section\",
\"article\",
\"aside\",
\"main\",
\"figure\",
\"figcaption\",
\"blockquote\",
\"q\",
\"hr\",
\"code\",
\"pre\",
\"iframe\",
\"video\",
\"audio\",
\"canvas\",
\"svg\",
\"path\",
\"circle\",
\"rect\",
\"line\",
\"polyline\",
\"polygon\",
\"g\",
\"use\",
\"symbol\"
]
},
\"label\": {
\"type\": \"string\",
\"description\": \"The label of the UI component, used for buttons or form fields\"
},
\"children\": {
\"type\": \"array\",
\"description\": \"Nested UI components\",
\"items\": {
\"$ref\": \"#\"
}
},
\"attributes\": {
\"type\": \"array\",
\"description\": \"Arbitrary attributes for the UI component, suitable for any element using Tailwind framework\",
\"items\": {
\"type\": \"object\",
\"properties\": {
\"name\": {
\"type\": \"string\",
\"description\": \"The name of the attribute, for example onClick or className\"
},
\"value\": {
\"type\": \"string\",
\"description\": \"The value of the attribute using the Tailwind framework classes\"
}
},
\"additionalProperties\": false,
\"required\": [\"name\", \"value\"]
}
}
},
\"required\": [\"type\", \"label\", \"children\", \"attributes\"],
\"additionalProperties\": false
}
}
}
}",
"sendBody": true,
"sendHeaders": true,
"specifyBody": "json",
"authentication": "predefinedCredentialType",
"headerParameters": {
"parameters": [
{
"name": "Content-Type",
"value": "application/json"
}
]
},
"nodeCredentialType": "openAiApi"
},
"credentials": {
"openAiApi": {
"id": "WqzqjezKh8VtxdqA",
"name": "OpenAi account - Baptiste"
}
},
"typeVersion": 4.2
},
{
"id": "24e5ca73-a3b3-4096-8c66-d84838d89b0c",
"name": "OpenAI - JSON to HTML",
"type": "@n8n/n8n-nodes-langchain.openAi",
"position": [
1420,
380
],
"parameters": {
"modelId": {
"__rl": true,
"mode": "list",
"value": "gpt-4o-mini",
"cachedResultName": "GPT-4O-MINI"
},
"options": {
"temperature": 0.2
},
"messages": {
"values": [
{
"role": "system",
"content": "You convert a JSON to HTML.
The JSON output has the following fields:
- html: the page HTML
- title: the page title"
},
{
"content": "={{ $json.choices[0].message.content }}"
}
]
},
"jsonOutput": true
},
"credentials": {
"openAiApi": {
"id": "WqzqjezKh8VtxdqA",
"name": "OpenAi account - Baptiste"
}
},
"typeVersion": 1.3
},
{
"id": "c50bdc84-ba59-4f30-acf7-496cee25068d",
"name": "Format the HTML result",
"type": "n8n-nodes-base.html",
"position": [
1940,
380
],
"parameters": {
"html": "<!DOCTYPE html>
<html>
<head>
<meta charset=\"UTF-8\" />
<script src=\"https://cdn.tailwindcss.com\"></script>
<title>{{ $json.message.content.title }}</title>
</head>
<body>
{{ $json.message.content.html }}
</body>
</html>"
},
"typeVersion": 1.2
},
{
"id": "193093f4-b1ce-4964-ab10-c3208e343c69",
"name": "Sticky Note",
"type": "n8n-nodes-base.stickyNote",
"position": [
1134,
62
],
"parameters": {
"color": 7,
"width": 638,
"height": 503,
"content": "## Generate HTML from user query
**HTTP Request node**
- Send the user query to OpenAI, with a defined JSON response format - *using HTTP Request node as it has not yet been implemented in the OpenAI nodes*
- The response format is inspired by the [Structured Output defined in OpenAI Introduction post](https://openai.com/index/introducing-structured-outputs-in-the-api)
- The output is a JSON containing HTML components and attributed
**OpenAI node**
- Format the response from the previous node from JSON format to HTML format"
},
"typeVersion": 1
},
{
"id": "0371156a-211f-4d92-82b1-f14fe60d4b6b",
"name": "Sticky Note1",
"type": "n8n-nodes-base.stickyNote",
"position": [
0,
60
],
"parameters": {
"color": 7,
"width": 768,
"height": 503,
"content": "## Workflow: Dynamically generate an HTML page from a user request using OpenAI Structured Output
**Overview**
- This workflow is a experiment to build HTML pages from a user input using the new Structured Output from OpenAI.
- The Structured Output could be used in a variety of cases. Essentially, it guarantees the output from the GPT will follow a defined structure (JSON object).
- It uses Tailwind CSS to make it slightly nicer, but any
**How it works**
- Once active, go to the production URL and add what you'd like to build as the parameter \"query\"
- Example: https://production_url.com?query=a%20signup%20form
- OpenAI nodes will first output the UI as a JSON then convert it to HTML
- Finally, the response is integrated in a HTML container and rendered to the user
**Further thoughts**
- Results are not yet amazing, it is hard to see the direct value of such an experiment
- But it showcase the potential of the Structured Output. Being able to guarantee the output format is key to build robust AI applications."
},
"typeVersion": 1
},
{
"id": "06380781-5189-4d99-9ecd-d8913ce40fd5",
"name": "Webhook",
"type": "n8n-nodes-base.webhook",
"position": [
820,
380
],
"webhookId": "d962c916-6369-431a-9d80-af6e6a50fdf5",
"parameters": {
"path": "d962c916-6369-431a-9d80-af6e6a50fdf5",
"options": {
"allowedOrigins": "*"
},
"responseMode": "responseNode"
},
"typeVersion": 2
}
],
"active": true,
"pinData": {},
"settings": {
"executionOrder": "v1"
},
"versionId": "d2307a2a-5427-4769-94a6-10eab703a788",
"connections": {
"Webhook": {
"main": [
[
{
"node": "Open AI - Using Structured Output",
"type": "main",
"index": 0
}
]
]
},
"OpenAI - JSON to HTML": {
"main": [
[
{
"node": "Format the HTML result",
"type": "main",
"index": 0
}
]
]
},
"Format the HTML result": {
"main": [
[
{
"node": "Respond to Webhook",
"type": "main",
"index": 0
}
]
]
},
"Open AI - Using Structured Output": {
"main": [
[
{
"node": "OpenAI - JSON to HTML",
"type": "main",
"index": 0
}
]
]
}
}
}
功能特点
- 自动检测新邮件
- AI智能内容分析
- 自定义分类规则
- 批量处理能力
- 详细的处理日志
技术分析
节点类型及作用
- Respondtowebhook
- Httprequest
- @N8N/N8N Nodes Langchain.Openai
- Html
- Stickynote
复杂度评估
配置难度:
维护难度:
扩展性:
实施指南
前置条件
- 有效的Gmail账户
- n8n平台访问权限
- Google API凭证
- AI分类服务订阅
配置步骤
- 在n8n中导入工作流JSON文件
- 配置Gmail节点的认证信息
- 设置AI分类器的API密钥
- 自定义分类规则和标签映射
- 测试工作流执行
- 配置定时触发器(可选)
关键参数
| 参数名称 | 默认值 | 说明 |
|---|---|---|
| maxEmails | 50 | 单次处理的最大邮件数量 |
| confidenceThreshold | 0.8 | 分类置信度阈值 |
| autoLabel | true | 是否自动添加标签 |
最佳实践
优化建议
- 定期更新AI分类模型以提高准确性
- 根据邮件量调整处理批次大小
- 设置合理的分类置信度阈值
- 定期清理过期的分类规则
安全注意事项
- 妥善保管API密钥和认证信息
- 限制工作流的访问权限
- 定期审查处理日志
- 启用双因素认证保护Gmail账户
性能优化
- 使用增量处理减少重复工作
- 缓存频繁访问的数据
- 并行处理多个邮件分类任务
- 监控系统资源使用情况
故障排除
常见问题
邮件未被正确分类
检查AI分类器的置信度阈值设置,适当降低阈值或更新训练数据。
Gmail认证失败
确认Google API凭证有效且具有正确的权限范围,重新进行OAuth授权。
调试技巧
- 启用详细日志记录查看每个步骤的执行情况
- 使用测试邮件验证分类逻辑
- 检查网络连接和API服务状态
- 逐步执行工作流定位问题节点
错误处理
工作流包含以下错误处理机制:
- 网络超时自动重试(最多3次)
- API错误记录和告警
- 处理失败邮件的隔离机制
- 异常情况下的回滚操作