Hébergement VPS n8n | Automatisation sécurisée et évolutive
Formation n8n en Français : Automatisation & Agents IA
This is the shooting course for rebuilding the workflow node by node, live, without ever opening the finished JSON. Everything below is copy-paste ready.
<aside> 🎯
What this workflow does: scrapes active Facebook ads for any keyword/niche via Apify's Ad Library Scraper, filters out low-engagement advertisers, routes each ad by format (video / image / text), analyzes it with GPT-4 Vision + an AI Agent (GPT-4.1), and logs a strategic summary + a rewritten, non-plagiarized ad copy to Google Sheets.
</aside>
<aside> 🆚
Improvement vs the original template: the original used Gemini + Google Drive to analyze video ads (upload video → Gemini upload session → poll → describe). This version removes Gemini and Google Drive entirely: it analyzes the video's preview cover frame with GPT-4 Vision instead, which is far simpler, cheaper, and needs only one credential (OpenAI).
</aside>
| Credential | Type | Where to get it |
|---|---|---|
| Apify API | Header Auth (Name: Authorization, Value: Bearer YOUR_APIFY_TOKEN) |
Sign up at apify.com → Settings → Integrations → copy your API token |
| OpenAi account | OpenAI API | platform.openai.com → API keys → create a new secret key (needs GPT-4o / GPT-4.1 access) |
| Google Sheets account | Google Sheets OAuth2 | Connect your Google account in n8n and grant Sheets access |
<aside> ⚠️
Make sure your Apify account has access to a Facebook Ad Library Scraper actor. This build uses actor XtaWFhbtfxyzqrFmd (run-sync-get-dataset-items endpoint) — swap the actor ID in the HTTP node's URL if you use a different one.
</aside>
Type: Manual Trigger (n8n-nodes-base.manualTrigger, v1)
No configuration needed. Click "Execute workflow" to run a test.