Hébergement VPS n8n | Automatisation sécurisée et évolutive
Formation n8n en Français : Automatisation & Agents IA
This course lets you rebuild the workflow from scratch, node by node, live. Everything you need to copy-paste (prompts, expressions, JSON bodies, URLs) is in code blocks below.
What it does: a Telegram bot receives a product photo, OpenAI Vision analyzes it, an AI agent writes a cinematic text-to-video prompt + a YouTube title & description, AtlasCloud Seedance 2.0 generates a vertical 9:16 video, and after two human approvals on Telegram, Blotato publishes it to YouTube as an AI-flagged Short.
Honesty note: the AI proposes, you check the quality before letting it run. Two approval gates keep you in control.
Telegram Trigger (photo)
-> Configuration (Set)
-> Download product photo (Telegram file:get, binary)
-> Analyze product photo (OpenAI Vision)
-> Generate video ad prompt (AI Agent + Model + Output Parser)
-> Review video prompt on Telegram (sendAndWait, double approval)
-> Prompt approved? (IF)
false -> Notify prompt declined (end)
true -> Notify generating video
-> Generate video (AtlasCloud, HTTP POST)
-> Wait for render
-> Check video status (AtlasCloud, HTTP GET)
-> Video ready? (IF)
false -> Wait for render (loop)
true -> Send video preview on Telegram
-> Review video on Telegram (sendAndWait)
-> Video approved? (IF)
false -> Notify video declined (end)
true -> Publish to YouTube (Blotato)
-> Notify published (end)