基于GPT 4.1 mini和Firecrawl的网络抓取和截图自动化
中级
这是一个Market Research, AI Summarization领域的自动化工作流,包含 13 个节点。主要使用 HttpRequest, Agent, HttpRequestTool, ChatTrigger, LmChatOpenRouter 等节点。 基于GPT 4.1 mini和Firecrawl的网络抓取和截图自动化
前置要求
- •可能需要目标 API 的认证凭证
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
{
"meta": {
"instanceId": "07969d022f78c86e07b98687c867015ba10dcaf713875d5ef0558ba6f30ac8fc"
},
"nodes": [
{
"id": "d324309d-a8e6-49a7-9aba-6539822ccc09",
"name": "Site",
"type": "n8n-nodes-base.httpRequest",
"position": [
3408,
1472
],
"parameters": {
"url": "https://api.firecrawl.dev/v1/search",
"method": "POST",
"options": {},
"jsonBody": "{\n \"query\": \"nate herk site:www.geeky-gadgets.com\",\n \"limit\": 5\n}",
"sendBody": true,
"sendHeaders": true,
"specifyBody": "json",
"headerParameters": {
"parameters": [
{}
]
}
},
"typeVersion": 4.2
},
{
"id": "0aea3a14-37ac-46c6-8e8c-a12c3750e801",
"name": "In URL",
"type": "n8n-nodes-base.httpRequest",
"position": [
3408,
1680
],
"parameters": {
"url": "https://api.firecrawl.dev/v1/search",
"method": "POST",
"options": {},
"jsonBody": "{\n \"query\": \"nate herk inurl:skool\",\n \"limit\": 5\n}",
"sendBody": true,
"sendHeaders": true,
"specifyBody": "json",
"headerParameters": {
"parameters": [
{}
]
}
},
"typeVersion": 4.2
},
{
"id": "f6c6becf-a6de-4523-846b-9ec51d485a2f",
"name": "Exclusion",
"type": "n8n-nodes-base.httpRequest",
"position": [
3760,
1472
],
"parameters": {
"url": "https://api.firecrawl.dev/v1/search",
"method": "POST",
"options": {},
"jsonBody": "{\n \"query\": \"nate herk -inurl:skool\",\n \"limit\": 6\n}",
"sendBody": true,
"sendHeaders": true,
"specifyBody": "json",
"headerParameters": {
"parameters": [
{}
]
}
},
"typeVersion": 4.2
},
{
"id": "8107f822-b505-4158-8c13-8f1bf7774db0",
"name": "Pro",
"type": "n8n-nodes-base.httpRequest",
"position": [
3760,
1680
],
"parameters": {
"url": "https://api.firecrawl.dev/v1/search",
"method": "POST",
"options": {},
"jsonBody": "{\n \"query\": \"Nate Herk site:youtube.com -shorts intitle:automation\",\n \"limit\": 5\n}",
"sendBody": true,
"sendHeaders": true,
"specifyBody": "json",
"headerParameters": {
"parameters": [
{}
]
}
},
"typeVersion": 4.2
},
{
"id": "b3aacfb1-b468-4674-99c0-9bf59147bfac",
"name": "Sticky Note1",
"type": "n8n-nodes-base.stickyNote",
"position": [
3280,
1408
],
"parameters": {
"color": 5,
"width": 340,
"height": 200,
"content": "## Within a Specific Website"
},
"typeVersion": 1
},
{
"id": "a6a4c52b-6388-4fba-832f-2f77c9c75d62",
"name": "Sticky Note2",
"type": "n8n-nodes-base.stickyNote",
"position": [
3280,
1632
],
"parameters": {
"color": 5,
"width": 340,
"height": 200,
"content": "## Word Appears in URL"
},
"typeVersion": 1
},
{
"id": "c535f40e-4dae-4fe7-bbd7-c87794e46fcc",
"name": "Sticky Note3",
"type": "n8n-nodes-base.stickyNote",
"position": [
3648,
1408
],
"parameters": {
"color": 5,
"width": 340,
"height": 200,
"content": "## Exclude Terms\n"
},
"typeVersion": 1
},
{
"id": "5bdedd5c-9068-41aa-8ccc-35ff6acd2360",
"name": "Sticky Note4",
"type": "n8n-nodes-base.stickyNote",
"position": [
3648,
1632
],
"parameters": {
"color": 5,
"width": 340,
"height": 200,
"content": "## Pro Tip"
},
"typeVersion": 1
},
{
"id": "e287e808-3690-46dc-a85b-aa0ceed922bd",
"name": "When chat message received",
"type": "@n8n/n8n-nodes-langchain.chatTrigger",
"position": [
3376,
1120
],
"webhookId": "5634d633-9287-41c9-ab57-bea1816b3b0a",
"parameters": {
"options": {}
},
"typeVersion": 1.1
},
{
"id": "d0a5315a-1d4a-41a7-b68c-ab31686370e1",
"name": "GPT 4.1 mini",
"type": "@n8n/n8n-nodes-langchain.lmChatOpenRouter",
"position": [
3488,
1264
],
"parameters": {
"options": {}
},
"typeVersion": 1
},
{
"id": "19753a83-34d3-4d61-a1ab-4d7d90f5245d",
"name": "Firecrawl Search",
"type": "n8n-nodes-base.httpRequestTool",
"position": [
3776,
1248
],
"parameters": {
"url": "https://api.firecrawl.dev/v1/search",
"method": "POST",
"options": {},
"jsonBody": "={\n \"query\": \"{{$fromAI(\"searchQuery\")}}\",\n \"limit\": {{$fromAI(\"limit\",\"the number of search results requested\", number)}},\n \"scrapeOptions\": {\n \"formats\": [\"markdown\", \"screenshot@fullPage\"]}\n}",
"sendBody": true,
"specifyBody": "json",
"toolDescription": "Use this tool to search the internet"
},
"typeVersion": 4.2
},
{
"id": "34fa76be-ef2a-4962-9ff1-9d5268525e5f",
"name": "Search Agent",
"type": "@n8n/n8n-nodes-langchain.agent",
"position": [
3536,
1120
],
"parameters": {
"options": {
"systemMessage": "=# Overview\nYou are a Firecrawl Search Query Generator Agent.\n\nYou have access to a tool called FireCrawl Search.\n\nYour job is to:\n1. Convert natural language instructions into a properly formatted Firecrawl query string using supported operators.\n2. Send that query — along with a `limit` parameter — to the **FireCrawl Search** tool.\n\n## Query Construction Rules:\n\nUse the following rules to transform user input:\n\n- If the user specifies an exact phrase \n → `Nate Herk`\n\n- To exclude terms or domains, prefix them with `-` \n → `-linkedin`, `-site:facebook.com`\n\n- If the user wants to search within a specific site, use `site:` \n → `site:youtube.com`\n\n- If the user wants a word in the URL, use `inurl:` \n → `inurl:nateherk`\n\n- If the user wants multiple words in the URL, use `allinurl:` \n → `allinurl:nate herk agent`\n\n- If the user wants a word in the title of the page, use `intitle:` \n → `intitle:automation`\n\n- If the user wants multiple words in the title, use `allintitle:` \n → `allintitle:ai agent tutorial`\n\n- If the user wants results related to a domain, use `related:` \n → `related:truehorizonai.com`\n\n\n## Tool Usage Instructions:\n\nAlways call the **FireCrawls Search** tool with:\n- `query`: the final constructed query string \n- `limit`: the number of results the user requested\n\nIf the user does **not specify a limit**, default to: \n→ `limit: 5`\n\n## Output\n\nAlways return all of the important information for each item that the tool gives you. The the title, the URL, the results, etc. \n\nFor each 'data' object that is returned in the results, output every field that lives in that object.\n\n## Examples:\n\n**Input:** Search the internet for Nate Herk \n**Action:** Call FireCrawls Search with: \n```json\n{\n \"query\": Nate Herk,\n \"limit\": 5\n}\n```\n\n**Input:** Find pages with AI agent in the title from YouTube, show me 10 \n**Action:** \n```json\n{\n \"query\": intitle:AI agent site:youtube.com,\n \"limit\": 10\n}\n```\n\n**Input:** Show me results that mention Nate Herk but exclude LinkedIn \n**Action:** \n```json\n{\n \"query\": Nate Herk -site:linkedin.com,\n \"limit\": 5\n}\n```\n"
}
},
"typeVersion": 2
},
{
"id": "459d94c8-028d-4bd0-bf7a-406148441c17",
"name": "Sticky Note8",
"type": "n8n-nodes-base.stickyNote",
"position": [
3280,
1040
],
"parameters": {
"color": 6,
"width": 700,
"height": 340,
"content": "## Firecrawl Agent"
},
"typeVersion": 1
}
],
"pinData": {},
"connections": {
"GPT 4.1 mini": {
"ai_languageModel": [
[
{
"node": "Search Agent",
"type": "ai_languageModel",
"index": 0
}
]
]
},
"Firecrawl Search": {
"ai_tool": [
[
{
"node": "Search Agent",
"type": "ai_tool",
"index": 0
}
]
]
},
"When chat message received": {
"main": [
[
{
"node": "Search Agent",
"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?
中级 - 市场调研, AI 摘要总结
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
财务报告工具
财务报告 AI:通过 OpenRouter + Perplexity 生成简洁的 SEC 10-K/10-Q 摘要
Set
Code
Markdown
+6
17 节点Ashish Kumar Swain
市场调研
YouTube 潜在客户生成:使用 Apify 和 Gemini AI 将评论转化为丰富潜在客户
YouTube 潜在客户生成:使用 Apify 和 Gemini AI 将评论转化为丰富潜在客户
Code
Http Request
Google Sheets
+9
22 节点Msaid Mohamed el hadi
潜在客户开发
使用 Bright Data API 和 AI 抓取分析 Google 广告并发送邮件报告
使用 Bright Data API 和 AI 抓取分析 Google 广告并发送邮件报告
Set
Code
Gmail
+15
45 节点Zacharia Kimotho
市场调研
基于AI的会议研究与每日议程(Google日历、Attio CRM和Slack)
基于AI的会议研究与每日议程:使用Google日历、Attio CRM和Slack
If
Set
Code
+15
30 节点Harry Siggins
AI 摘要总结
RAG 重排序
使用 Supabase、OpenAI 和 Cohere 重排序器从文档中回答问题
Code
Google Drive
Manual Trigger
+9
26 节点Luan Correia
内部知识库
自动化会议准备
使用 GPT-5 和 Gemini 研究从日历到 Slack 通过 Attio CRM 自动准备会议
If
Set
Code
+16
39 节点Harry Siggins
AI 摘要总结