Trigger: Listens for new images sent to my Telegram bot.
Fetch: Downloads the image file from Telegram using its file_id.
Convert: Turns the image into Base64 so it’s ready for AI processing.
Describe: Sends the Base64 image to GROQ LLaMA to get a detailed description.
Extract: Pulls only the text from the AI’s response.
Detect & Solve: AI Agent scans the text for math problems, solves them, and creates a custom Veo video prompt.
Format: Structures the math problem, answer, and video prompt in clean JSON.
Setup: Stores API keys, model details, and the generated prompt for Veo.
Generate Video: Sends the prompt to Google Vertex AI Veo to create the teacher video.
Wait: Pauses 1 minute for video generation to finish.
Fetch Video: Retrieves the completed video from Veo.
Convert: Turns the Base64 video into an .mp4 file.
Deliver: Sends the final AI-generated teacher video back to the same Telegram chat.
Note: You can also add WhatsApp node instead of Telegram.
Connect with me on IG if you want to build something similar in n8n.
Cheers 🥂
Harshad