Robot de renseignement de marché piloté par Firecrawl AI : livraison automatisée des insights de nouvelles

• •Token Bot Slack ou URL Webhook
• •Peut nécessiter les informations d'identification d'authentification de l'API cible
• •Clé API OpenAI

{
  "id": "Ut6jbfsK7IvhGJVL",
  "meta": {
    "instanceId": "87e85f497d2a9f18665002b06f3ba96bde7b80557a792319a7373c9549c8959f",
    "templateId": "4588",
    "templateCredsSetupCompleted": true
  },
  "name": "Firecrawl AI-Powered Market Intelligence Bot: Automated News Insights Delivery",
  "tags": [],
  "nodes": [
    {
      "id": "471044c1-cafd-4810-973d-b40c74ef6999",
      "name": "Déclencheur de Recherche Marchande Quotidienne",
      "type": "n8n-nodes-base.scheduleTrigger",
      "position": [
        2512,
        784
      ],
      "parameters": {
        "rule": {
          "interval": [
            {
              "triggerAtHour": 8
            }
          ]
        }
      },
      "typeVersion": 1.2
    },
    {
      "id": "8f6e97b6-6910-4c6b-8c9a-29a0cf95ac09",
      "name": "Explorer TechCrunch (FireCrawl)",
      "type": "n8n-nodes-base.httpRequest",
      "position": [
        2704,
        784
      ],
      "parameters": {
        "url": "https://api.firecrawl.dev/v1/crawl",
        "method": "POST",
        "options": {},
        "jsonBody": "{\n  \"url\": \"https://techcrunch.com\",\n  \"limit\": 20,\n  \"includePaths\": [\"2025/\"],\n  \"scrapeOptions\": {\n    \"formats\": [\"markdown\"],\n    \"onlyMainContent\": true,\n    \"parsePDF\": true,\n    \"maxAge\": 14400000\n  }\n}",
        "sendBody": true,
        "specifyBody": "json",
        "authentication": "genericCredentialType",
        "genericAuthType": "httpBearerAuth"
      },
      "typeVersion": 4.2
    },
    {
      "id": "a326e8fd-bfc8-4380-9c66-20481aaa8a6c",
      "name": "Agent de Synthèse",
      "type": "@n8n/n8n-nodes-langchain.agent",
      "position": [
        3328,
        784
      ],
      "parameters": {
        "text": "=You are an AI research assistant. First, determine if this article is related to artificial intelligence, machine learning, AI companies, or AI technology.\n\nIf the article IS AI-related, provide a summary in 3 bullet points.\nIf the article is NOT AI-related, respond with exactly: \"NOT_AI_RELATED\"\n\nArticle details:\nTitle: {{ $json.title }}\nDescription: {{ $json.description }}\nContent: {{ $json.content }}\n\nFormat for AI articles:\n{{ $json.title }}\n\nSummary:\n- [Bullet point 1]\n- [Bullet point 2] \n- [Bullet point 3]\n\nLink: {{ $json.url }}",
        "options": {},
        "promptType": "define"
      },
      "typeVersion": 1.9
    },
    {
      "id": "d8576228-6261-4654-84de-0864e398c22d",
      "name": "OpenAI Summarizer",
      "type": "@n8n/n8n-nodes-langchain.lmChatOpenAi",
      "position": [
        3328,
        1008
      ],
      "parameters": {
        "model": {
          "__rl": true,
          "mode": "list",
          "value": "gpt-4o-mini"
        },
        "options": {}
      },
      "credentials": {
        "openAiApi": {
          "id": "byfe88CyXUHFOXUZ",
          "name": "OpenAi account"
        }
      },
      "typeVersion": 1.2
    },
    {
      "id": "f5965b4a-4ab1-4e5f-868c-2dc11c59fc28",
      "name": "Envoyer le Résumé à Slack",
      "type": "n8n-nodes-base.slack",
      "position": [
        3792,
        784
      ],
      "webhookId": "c2e00b18-42bd-49b7-bc4c-05d60633a7c8",
      "parameters": {
        "text": "=🔍 AI Research Summary:\n{{ $json.output }}",
        "select": "channel",
        "channelId": {
          "__rl": true,
          "mode": "name",
          "value": "#general"
        },
        "otherOptions": {},
        "authentication": "oAuth2"
      },
      "executeOnce": false,
      "typeVersion": 2.3
    },
    {
      "id": "fb908a45-0faf-4068-9420-8046ab4a8d36",
      "name": "Attendre",
      "type": "n8n-nodes-base.wait",
      "position": [
        2880,
        784
      ],
      "webhookId": "b210f38a-da89-4bc5-a1e6-c80f9b89a426",
      "parameters": {
        "amount": 60
      },
      "typeVersion": 1.1
    },
    {
      "id": "a89a5bc1-5f76-40b5-aca3-adc380fd4c0d",
      "name": "Recevoir les Résultats Firecrawl",
      "type": "n8n-nodes-base.httpRequest",
      "position": [
        3024,
        784
      ],
      "parameters": {
        "url": "=https://api.firecrawl.dev/v1/crawl/{{$json.id}}",
        "options": {},
        "authentication": "genericCredentialType",
        "genericAuthType": "httpBearerAuth"
      },
      "credentials": {
        "httpBearerAuth": {
          "id": "U4MsswmPveLVhYYR",
          "name": "Bearer Auth account 4"
        }
      },
      "typeVersion": 4.2
    },
    {
      "id": "5ddeaef9-cd9d-4f55-b82d-1edfce6429df",
      "name": "Diviser la Sortie",
      "type": "n8n-nodes-base.code",
      "position": [
        3184,
        784
      ],
      "parameters": {
        "jsCode": "// Split crawled articles into individual items for processing\nif (!$json.data || $json.data.length === 0) {\n  console.log(\"No data available\");\n  return [];\n}\n\nconsole.log(`Processing ${$json.data.length} articles individually`);\n\n// Return each article as a separate n8n item\nreturn $json.data.map(article => ({\n  json: {\n    title: article.metadata?.title || 'No title',\n    url: article.sourceURL || '',\n    content: (article.markdown || article.content || '').substring(0, 1000),\n    description: article.metadata?.description || '',\n    publishDate: article.metadata?.publishDate || ''\n  }\n}));"
      },
      "typeVersion": 2
    },
    {
      "id": "c8cc970e-313e-45c9-bad5-9dd32ccbfee3",
      "name": "Filtrer les Messages",
      "type": "n8n-nodes-base.code",
      "position": [
        3616,
        784
      ],
      "parameters": {
        "jsCode": "// Process all items at once and filter out NOT_AI_RELATED\nconst filteredItems = [];\n\n$input.all().forEach(item => {\n  if (item.json.output && item.json.output.trim() !== 'NOT_AI_RELATED') {\n    filteredItems.push(item);\n  } else {\n    console.log('Filtered out non-AI article');\n  }\n});\n\nconsole.log(`Passing ${filteredItems.length} AI articles to Slack`);\nreturn filteredItems;"
      },
      "typeVersion": 2
    },
    {
      "id": "51e047ef-791d-46cd-abcd-98c2b55104cd",
      "name": "Note Adhésive",
      "type": "n8n-nodes-base.stickyNote",
      "position": [
        2512,
        -384
      ],
      "parameters": {
        "width": 944,
        "height": 1056,
        "content": "# AI-Powered Market Intelligence Bot\n\n## Overview\nThis n8n workflow automatically monitors TechCrunch for AI-related articles, using Firecrawl's AI-powered web scraping to extract clean, structured data and OpenAI to generate actionable summaries delivered directly to your team's Slack channel.\n\n## 🚀 1. Trigger & Scheduling\n\n### Daily Method Research Trigger\n- Initiates automated TechCrunch monitoring on a scheduled basis\n- Default: Daily execution to capture fresh AI industry news and trends\n- Configurable for different frequencies based on market velocity\n- Ensures consistent intelligence gathering without manual intervention\n\n## 🔍 2. Web Scraping & Data Collection\n\n### Firecrawl Integration (HTTP Request Node)\n- Targets TechCrunch specifically: `https://techcrunch.com` (though, can obviously be changed)\n- Crawls a specificed amount of articles with `\"limit\" -- by default this is set to 20, otherwise it would keep scraping. You can remove this limit to scrape all. \n- Focuses on current year content with `\"includePaths\": [\"2025/\"]` In 2026, you'll need to update this.\n- Extracts markdown format for clean, structured content\n- Uses `\"onlyMainContent\": true` to avoid ads and navigation clutter\n- Built-in anti-bot detection bypass and automatic JavaScript rendering\n\n### Code Text/Concat Processing\n- Filters scraped TechCrunch articles for AI relevance\n- AI research assistant determines if articles relate to artificial intelligence, machine learning, AI companies, or AI technology\n- Automatically excludes non-AI content with \"NOT_AI_RELATED\" filtering\n- Processes article titles, descriptions, and full content for comprehensive analysis\n\n## ⏱️ 3. Processing Flow Control\n\n### Wait Node\n- Implements controlled delays to respect TechCrunch's server resources\n- Enough time to receive results back"
      },
      "typeVersion": 1
    },
    {
      "id": "92923278-85b3-4def-a7e2-8967e55f9004",
      "name": "Note Adhésive1",
      "type": "n8n-nodes-base.stickyNote",
      "position": [
        4384,
        -352
      ],
      "parameters": {
        "color": 4,
        "width": 1200,
        "height": 1184,
        "content": "## 🔧 Configuration & Customization\n\n### TechCrunch Scraping Setup\n- Configured to crawl `https://techcrunch.com` with 20 article limit\n- Focuses on current year content with date path filtering\n- Markdown extraction for clean, readable content processing\n- 4-hour cache setting (`\"maxAge\": 14400000`) for efficiency\n\n### AI Relevance Detection\n- Built-in AI research assistant determines article relevance\n- Filters specifically for artificial intelligence, machine learning, AI companies, and AI technology\n- Binary classification: AI-related articles get summarized, others get filtered out\n- Reduces noise by focusing only on AI industry developments\n\n### Slack Integration\n- TechCrunch-specific formatting with source attribution\n- AI-focused message structure with bullet point summaries\n- Integration with Slack threads for AI trend discussions\n- Notification preferences for AI industry updates\n\n## 💡 Best Practices\n\n1. **Monitor TechCrunch Coverage**: Track how much AI content TechCrunch publishes daily\n2. **AI Relevance Accuracy**: Review filtered articles to ensure important AI stories aren't missed\n3. **Summary Quality**: Periodically check that 3-bullet summaries capture key AI insights\n4. **Team Engagement**: Use Slack reactions to identify most valuable AI trends\n5. **Source Verification**: Always include TechCrunch links for fact-checking and deeper reading\n\n## 🔄 Maintenance & Monitoring\n\n- **TechCrunch Access**: Monitor Firecrawl success rates for TechCrunch crawling\n- **AI Classification**: Track accuracy of AI relevance detection\n- **Content Freshness**: Ensure date filtering captures latest AI articles\n- **Delivery Success**: Confirm AI summaries reach Slack channels consistently\n- **Cost Management**: Monitor OpenAI usage for TechCrunch article processing\n\n## 🎯 Business Impact\n\nThis workflow transforms AI market intelligence from manual TechCrunch reading into an automated pipeline that delivers:\n\n- **Daily AI-focused insights** from TechCrunch without manual browsing\n- **Filtered relevance** ensuring only AI industry news reaches your team\n- **Structured summaries** optimized for quick consumption of AI trends\n- **Team awareness** of AI developments through integrated Slack discussions\n- **Source credibility** with direct links to TechCrunch articles\n\nPerfect for AI product teams, tech investors, and executives who need to stay current on artificial intelligence developments without spending time manually browsing TechCrunch."
      },
      "typeVersion": 1
    },
    {
      "id": "52f22863-78b7-4f12-9d92-5cfe81b4c8c9",
      "name": "Note Adhésive2",
      "type": "n8n-nodes-base.stickyNote",
      "position": [
        3600,
        -352
      ],
      "parameters": {
        "color": 5,
        "width": 736,
        "height": 1056,
        "content": "## 🤖 4. AI-Powered Analysis & Summarization\n\n### OpenAI Summarizer (AI Agent)\n- Uses GPT-4o-mini for cost-effective, high-quality analysis\n- AI research assistant prompt determines AI relevance before summarizing\n- For AI-related articles: generates 3 bullet point summaries\n- For non-AI articles: responds with \"NOT_AI_RELATED\" to filter out\n- Processes TechCrunch article title, description, and full content\n- Maintains source attribution with original TechCrunch URLs\n\n### Split Output Processing\n- Distributes processed TechCrunch articles across parallel analysis streams\n- Enables simultaneous processing of multiple articles\n- Optimizes workflow performance for large news batches\n- Maintains data integrity during multi-path processing\n\n## 📊 5. Intelligence Filtering & Quality Control\n\n### Filter Messages Node\n- Removes articles marked as \"NOT_AI_RELATED\" by the AI assistant\n- Applies relevance scoring to prevent information overload\n- Ensures only AI-focused TechCrunch content reaches stakeholders\n- Customizable filtering logic based on AI relevance confidence\n\n## 📤 6. Delivery & Distribution\n\n### Send Summary to Slack\n- Delivers formatted AI intelligence from TechCrunch to specified channels\n- Message format includes: Article title + 3 bullet point summary + TechCrunch source link\n- Supports threaded conversations for team discussions about AI trends\n- Click-through links to original TechCrunch articles for deeper research\n- Rich formatting optimized for mobile and desktop consumptio"
      },
      "typeVersion": 1
    }
  ],
  "active": false,
  "pinData": {},
  "settings": {},
  "versionId": "1a74e487-1f09-4046-8c1c-149d0c748911",
  "connections": {
    "fb908a45-0faf-4068-9420-8046ab4a8d36": {
      "main": [
        [
          {
            "node": "a89a5bc1-5f76-40b5-aca3-adc380fd4c0d",
            "type": "main",
            "index": 0
          }
        ]
      ]
    },
    "5ddeaef9-cd9d-4f55-b82d-1edfce6429df": {
      "main": [
        [
          {
            "node": "a326e8fd-bfc8-4380-9c66-20481aaa8a6c",
            "type": "main",
            "index": 0
          }
        ]
      ]
    },
    "c8cc970e-313e-45c9-bad5-9dd32ccbfee3": {
      "main": [
        [
          {
            "node": "f5965b4a-4ab1-4e5f-868c-2dc11c59fc28",
            "type": "main",
            "index": 0
          }
        ]
      ]
    },
    "a326e8fd-bfc8-4380-9c66-20481aaa8a6c": {
      "main": [
        [
          {
            "node": "c8cc970e-313e-45c9-bad5-9dd32ccbfee3",
            "type": "main",
            "index": 0
          }
        ]
      ]
    },
    "d8576228-6261-4654-84de-0864e398c22d": {
      "ai_languageModel": [
        [
          {
            "node": "a326e8fd-bfc8-4380-9c66-20481aaa8a6c",
            "type": "ai_languageModel",
            "index": 0
          }
        ]
      ]
    },
    "a89a5bc1-5f76-40b5-aca3-adc380fd4c0d": {
      "main": [
        [
          {
            "node": "5ddeaef9-cd9d-4f55-b82d-1edfce6429df",
            "type": "main",
            "index": 0
          }
        ]
      ]
    },
    "8f6e97b6-6910-4c6b-8c9a-29a0cf95ac09": {
      "main": [
        [
          {
            "node": "fb908a45-0faf-4068-9420-8046ab4a8d36",
            "type": "main",
            "index": 0
          }
        ]
      ]
    },
    "471044c1-cafd-4810-973d-b40c74ef6999": {
      "main": [
        [
          {
            "node": "8f6e97b6-6910-4c6b-8c9a-29a0cf95ac09",
            "type": "main",
            "index": 0
          }
        ]
      ]
    }
  }
}