使用Super RAG和GPT-5从知识源创建基于事实的文章
高级
这是一个AI RAG, Multimodal AI领域的自动化工作流，包含 19 个节点。主要使用 Set, SplitOut, Aggregate, FormTrigger, HttpRequest 等节点。 使用Super RAG和GPT-5从知识源创建基于事实的文章
前置要求
- •可能需要目标 API 的认证凭证
- •OpenAI API Key
使用的节点 (19)
工作流预览
可视化展示节点连接关系，支持缩放和平移
导出工作流
复制以下 JSON 配置到 n8n 导入，即可使用此工作流
{
"nodes": [
{
"id": "9a0f0715-0f7f-4779-8f3a-c35bcf2ca175",
"name": "新内容 - 生成研究问题",
"type": "@n8n/n8n-nodes-langchain.chainLlm",
"position": [
-1008,
528
],
"parameters": {
"text": "=Content title: {{ $json.Title }}\n\nArticle guidelines: {{ $json.Guidelines }}\n\n",
"messages": {
"messageValues": [
{
"message": "=You will receive a content title and an angle. Return 5–8 non-overlapping questions in JSON array format that cover everything needed to write excellent content as it breaks down the topic into sub-questions.\n\nGuidelines: \n- Start with simple, short broad questions for example to define the terms (e.g., What is X?, Why is X important?, How to do X?). \n- Then move into more specific, advanced, or analytical questions. \n- Ensure questions together form a complete coverage of the topic. \n\n## Output format:\n\nYou'll return the questions in such a JSON ARRAY:\n\n[\n {\n \"question\": \"Lorem ipsum dolor sit amet, consectetur adipiscing elit?\"\n },\n {\n \"question\": \"Lorem ipsum dolor sit amet, consectetur adipiscing elit, sed do eiusmod tempor incididunt ut labore et dolore magna aliqua?\"\n },\n {\n \"question\": \"Lorem ipsum dolor sit amet?\"\n },\n {\n \"question\": \"Lorem ipsum dolor sit amet, consectetur adipiscing elit, sed do eiusmod tempor incididunt ut labore et dolore magna aliqua, ut enim ad minim veniam?\"\n },\n {\n \"question\": \"Lorem ipsum dolor sit amet, consectetur?\"\n }\n]"
}
]
},
"promptType": "define",
"hasOutputParser": true
},
"typeVersion": 1.5
},
{
"id": "97aa9da3-f03f-4ef5-8645-6ca9a941dfd4",
"name": "格式化问题和答案",
"type": "n8n-nodes-base.set",
"position": [
320,
768
],
"parameters": {
"options": {},
"assignments": {
"assignments": [
{
"id": "1e33a1f4-a1a2-4962-ac72-cc518d7ff043",
"name": "Question",
"type": "string",
"value": "={{ $('Loop Over Questions').item.json.question }}"
},
{
"id": "903bcf38-13dd-48fb-8eb3-83f7a232aa53",
"name": "Answer",
"type": "string",
"value": "={{ $json.answer.replace(/\\{\\[(\\d+)\\]\\((.*?)\\)\\}/g, '([source]($2))') }}"
}
]
}
},
"typeVersion": 3.4
},
{
"id": "8ce76a1b-d79b-43e7-8b47-c61b6167b1af",
"name": "新内容 - 生成 AI 输出",
"type": "@n8n/n8n-nodes-langchain.chainLlm",
"position": [
144,
-48
],
"parameters": {
"text": "=Article title:\n\n{{ $('Prepare form values').first().json.Title }}\n\nArticle guidelines:\n\n{{ $('Prepare form values').first().json.Guidelines }}\n\n\nContent to leverage:\n\nThis Q&A research provides high-quality knowledge, insights, and sources for your content. Be sure to include source links in your output whenever a source was used.\n\n{{ JSON.stringify($json['questions and answers'], null, 2) }}\n",
"messages": {
"messageValues": [
{
"message": "=# Role\n\nYour role is to write an article based on the request in the user message.\n\n# What the user message contains\n\nThe user message includes the article title, any guidelines to follow, and comprehensive research material. This research is the sole basis for your article — do not invent information beyond it. When the research includes source links, integrate them smoothly as hyperlinks in the article.\n\n# How to write good articles\n\nYou excel at writing articles by making sure that they deliver value, are concise, seem like they are human-written, not using typical AI useless sentence formulations.\n\n# Your output format\n\nOutput only the full article.\n\n* Begin with a `# H1` title.\n* Use subheadings throughout the article."
}
]
},
"promptType": "define"
},
"typeVersion": 1.5
},
{
"id": "97939f7a-faf8-4b86-a566-ac55a1f3272a",
"name": "新文章表单",
"type": "n8n-nodes-base.formTrigger",
"position": [
-1568,
528
],
"webhookId": "61cdfeab-f3ce-4b9a-925b-63f813c267f9",
"parameters": {
"options": {},
"formTitle": "New article",
"formFields": {
"values": [
{
"fieldLabel": "Article title",
"placeholder": "10 ways to do Influencer Marketing in 2025",
"requiredField": true
},
{
"fieldLabel": "Article guidelines",
"placeholder": "Promote xyz and write in British English...",
"requiredField": true
}
]
},
"formDescription": "Fill in this form to trigger the generation of a new article."
},
"typeVersion": 2.3
},
{
"id": "dda874a4-b51e-46f1-b99b-a181beaae73c",
"name": "准备表单值",
"type": "n8n-nodes-base.set",
"position": [
-1312,
528
],
"parameters": {
"options": {},
"assignments": {
"assignments": [
{
"id": "ec4734ed-654f-478a-ab90-91bfcee1e208",
"name": "Title",
"type": "string",
"value": "={{ $json['Article title'] }}"
},
{
"id": "c034402e-a7b9-4c91-aaed-f24a838c3d91",
"name": "Guidelines",
"type": "string",
"value": "={{ $json['Article guidelines'] }}"
}
]
}
},
"typeVersion": 3.4
},
{
"id": "86265d5e-ad97-429c-af6b-3876b8d14a83",
"name": "结构化输出解析器",
"type": "@n8n/n8n-nodes-langchain.outputParserStructured",
"position": [
-880,
752
],
"parameters": {
"jsonSchemaExample": "[\n {\n \"question\": \"Lorem ipsum dolor sit amet, consectetur adipiscing elit?\"\n },\n {\n \"question\": \"Lorem ipsum dolor sit amet, consectetur adipiscing elit, sed do eiusmod tempor incididunt ut labore et dolore magna aliqua?\"\n },\n {\n \"question\": \"Lorem ipsum dolor sit amet?\"\n },\n {\n \"question\": \"Lorem ipsum dolor sit amet, consectetur adipiscing elit, sed do eiusmod tempor incididunt ut labore et dolore magna aliqua, ut enim ad minim veniam?\"\n },\n {\n \"question\": \"Lorem ipsum dolor sit amet, consectetur?\"\n }\n]"
},
"typeVersion": 1.2
},
{
"id": "41053ca5-776c-4f11-84ee-9799a44546f6",
"name": "GPT 5 mini",
"type": "@n8n/n8n-nodes-langchain.lmChatOpenAi",
"position": [
-1008,
752
],
"parameters": {
"model": {
"__rl": true,
"mode": "list",
"value": "gpt-5-mini",
"cachedResultName": "gpt-5-mini"
},
"options": {}
},
"typeVersion": 1.2
},
{
"id": "a41214eb-8650-43ab-9a29-26709c89e838",
"name": "拆分问题",
"type": "n8n-nodes-base.splitOut",
"position": [
-608,
528
],
"parameters": {
"options": {},
"fieldToSplitOut": "output"
},
"typeVersion": 1
},
{
"id": "fe8a86db-6972-4416-8223-ae7b9bc3c90f",
"name": "循环处理问题",
"type": "n8n-nodes-base.splitInBatches",
"position": [
-352,
512
],
"parameters": {
"options": {}
},
"typeVersion": 3
},
{
"id": "d6d46dbf-c20f-4c0d-a9f8-c3d3059d0678",
"name": "查询 Super Assistant",
"type": "n8n-nodes-base.httpRequest",
"position": [
-32,
544
],
"parameters": {
"url": "https://api.super.work/v1/super",
"method": "POST",
"options": {},
"sendBody": true,
"bodyParameters": {
"parameters": [
{
"name": "question",
"value": "={{ $json.question }}"
},
{
"name": "assistantId",
"value": "YOUR-ASSISTANT-ID"
}
]
}
},
"typeVersion": 4.2
},
{
"id": "1e6c7b70-60e0-44f9-a418-82b806b8d9ae",
"name": "GPT 5 chat",
"type": "@n8n/n8n-nodes-langchain.lmChatOpenAi",
"position": [
224,
176
],
"parameters": {
"model": {
"__rl": true,
"mode": "list",
"value": "gpt-5-chat-latest",
"cachedResultName": "gpt-5-chat-latest"
},
"options": {}
},
"typeVersion": 1.2
},
{
"id": "8b34795b-bc61-4c27-88dd-8a942c50a44a",
"name": "文章结果",
"type": "n8n-nodes-base.set",
"position": [
560,
-48
],
"parameters": {
"options": {},
"assignments": {
"assignments": [
{
"id": "d3b8c4bc-27d9-4d57-b8d6-3a40b84d7b7d",
"name": "Article",
"type": "string",
"value": "={{ $json.text }}"
}
]
}
},
"typeVersion": 3.4
},
{
"id": "5afd58d4-99f9-48c2-9d34-3d353cbbfc6f",
"name": "便签",
"type": "n8n-nodes-base.stickyNote",
"position": [
-96,
512
],
"parameters": {
"color": 7,
"width": 224,
"height": 320,
"content": ""
},
"typeVersion": 1
},
{
"id": "5429c604-cf0e-4635-b47e-7f436a5399c7",
"name": "聚合研究内容",
"type": "n8n-nodes-base.aggregate",
"position": [
-112,
160
],
"parameters": {
"options": {},
"aggregate": "aggregateAllItemData",
"destinationFieldName": "Content to leverage"
},
"typeVersion": 1
},
{
"id": "b86d6f11-dd1a-4b65-8342-6420b8ee4b61",
"name": "便签1",
"type": "n8n-nodes-base.stickyNote",
"position": [
-1152,
-256
],
"parameters": {
"width": 832,
"height": 560,
"content": "# 基于您知识库的 AI 文章撰写器"
},
"typeVersion": 1
},
{
"id": "615bc59d-5d05-4986-9bee-5a8ac089a390",
"name": "便签2",
"type": "n8n-nodes-base.stickyNote",
"position": [
-448,
384
],
"parameters": {
"color": 6,
"width": 976,
"height": 608,
"content": "## 使用 Super 逐一回答每个子问题"
},
"typeVersion": 1
},
{
"id": "a256f168-dbac-4a99-9617-32641916fae2",
"name": "便签3",
"type": "n8n-nodes-base.stickyNote",
"position": [
80,
-208
],
"parameters": {
"color": 5,
"width": 368,
"height": 512,
"content": "## AI 步骤：基于研究和初始请求撰写最终文章"
},
"typeVersion": 1
},
{
"id": "b909335e-7b65-41e8-b716-65605a278d29",
"name": "便签4",
"type": "n8n-nodes-base.stickyNote",
"position": [
-1152,
384
],
"parameters": {
"color": 6,
"width": 480,
"height": 608,
"content": "## 将主题分解为子问题"
},
"typeVersion": 1
},
{
"id": "b0e02f1f-73b2-4834-8a0e-2251ac1f53e1",
"name": "便签5",
"type": "n8n-nodes-base.stickyNote",
"position": [
-1648,
400
],
"parameters": {
"color": 4,
"width": 272,
"height": 304,
"content": "## 填写此表单以请求新文章"
},
"typeVersion": 1
}
],
"connections": {
"GPT 5 chat": {
"ai_languageModel": [
[
{
"node": "New content - Generate the AI output",
"type": "ai_languageModel",
"index": 0
}
]
]
},
"GPT 5 mini": {
"ai_languageModel": [
[
{
"node": "New content - generate research questions",
"type": "ai_languageModel",
"index": 0
}
]
]
},
"New article form": {
"main": [
[
{
"node": "Prepare form values",
"type": "main",
"index": 0
}
]
]
},
"Loop Over Questions": {
"main": [
[
{
"node": "Aggregate research content",
"type": "main",
"index": 0
}
],
[
{
"node": "Query Super Assistant",
"type": "main",
"index": 0
}
]
]
},
"Prepare form values": {
"main": [
[
{
"node": "New content - generate research questions",
"type": "main",
"index": 0
}
]
]
},
"Split Out Questions": {
"main": [
[
{
"node": "Loop Over Questions",
"type": "main",
"index": 0
}
]
]
},
"Query Super Assistant": {
"main": [
[
{
"node": "Format question and answer",
"type": "main",
"index": 0
}
]
]
},
"Structured Output Parser": {
"ai_outputParser": [
[
{
"node": "New content - generate research questions",
"type": "ai_outputParser",
"index": 0
}
]
]
},
"Aggregate research content": {
"main": [
[
{
"node": "New content - Generate the AI output",
"type": "main",
"index": 0
}
]
]
},
"Format question and answer": {
"main": [
[
{
"node": "Loop Over Questions",
"type": "main",
"index": 0
}
]
]
},
"New content - Generate the AI output": {
"main": [
[
{
"node": "Article result",
"type": "main",
"index": 0
}
]
]
},
"New content - generate research questions": {
"main": [
[
{
"node": "Split Out Questions",
"type": "main",
"index": 0
}
]
]
}
}
}常见问题
如何使用这个工作流？
复制上方的 JSON 配置代码，在您的 n8n 实例中创建新工作流并选择「从 JSON 导入」，粘贴配置后根据需要修改凭证设置即可。
这个工作流适合什么场景？
高级 - AI RAG 检索增强, 多模态 AI
需要付费吗？
本工作流完全免费，您可以直接导入使用。但请注意，工作流中使用的第三方服务（如 OpenAI API）可能需要您自行付费。
相关工作流推荐
使用AI规划、Linkup搜索和GPT-5创建研究支持的文章
使用AI规划、Linkup搜索和GPT-5创建研究支持的文章
Set
Split Out
Aggregate
+7
19 节点Guillaume Duvernay
内容创作
使用Lookio和OpenAI GPT从知识源创建基于事实的文章
使用Lookio和OpenAI GPT从知识源创建基于事实的文章
Set
Split Out
Aggregate
+7
19 节点Guillaume Duvernay
内容创作
使用Lookio、Linkup和GPT-5创建双源专家文章
使用Lookio、Linkup和GPT-5创建双源专家文章
Set
Split Out
Aggregate
+7
21 节点Guillaume Duvernay
动态AI网络研究员：从纯文本到自定义CSV
使用GPT-4和Linkup将纯文本转换为自定义CSV的动态AI网络研究员
Set
Code
Split Out
+7
16 节点Guillaume Duvernay
杂项
上下文混合RAG AI文案
Google Drive到Supabase上下文向量数据库同步用于RAG应用
If
Set
Code
+25
76 节点Michael Taleb
AI RAG 检索增强
使用GPT-5 Mini、Jira和表单界面自动化敏捷项目设置
使用GPT-5 Mini、Jira和表单界面自动化敏捷项目设置
Set
Jira
Gmail
+15
42 节点Billy Christi
多模态 AI
