Create AI UGC Videos from Telegram Voice Notes using OpenAI & HeyGen

Description

Turn Telegram voice messages into AI-generated videos and send the video link back to the chat. This workflow listens for messages, downloads the file, transcribes the audio, generates a video via HeyGen, and returns the final URL. Fast, automated UGC creation right inside Telegram. ✨

What This Template Does

• Listens for new Telegram messages and triggers the flow.
• Downloads attached files (e.g., voice memos).
• Transcribes the voice memo to text.
• Generates a video via HeyGen using the transcription.
• Polls HeyGen for video status, waits with a short buffer, and sends the final video URL back.

Key Benefits

• Automates UGC creation from Telegram messages ⚡
• Reduces manual steps from file handling to final video delivery 🧩
• Keeps users in Telegram for a smooth, chat-native experience 📲
• Modular nodes make it easy to customize or extend 🔧

Features

• Telegram message trigger and message updates 🛎️
• File download and retrieval for processing 📥
• Voice transcription step for audio-to-text 🎙️➡️📝
• HeyGen video generation via POST and status checks via GET 🌐
• 10-second buffer to allow external processing ⏱️
• Automatic return of the video URL to the chat 🔗

Requirements

• n8n instance (cloud or self-hosted) ⚙️
• Telegram Bot credentials configured in n8n (to receive messages and send replies) 🤖
• HeyGen API key for:
• POST https://api.heygen.com/v2/video/generate
• GET https://api.heygen.com/v1/video_status.get
• A transcription step configured in n8n for converting voice memos to text (node present in this workflow) 🗣📝

Target Audience

• Teams creating AI UGC directly from chat workflows 📢
• Marketing and social teams capturing quick briefs via voice 🧑
• Agencies standardizing UGC production pipelines 🏢
• Ops and support teams needing rapid content responses 🚀

Step-by-Step Setup Instructions (Concise)

• Connect Telegram Bot credentials in n8n and select the chat/channel for Message Trigger 🤝
• Add your HeyGen API key in n8n Credentials and map it in the HTTP Request nodes 🔑
• Ensure the transcription node is configured to convert audio to text and map output to video generation 🧭
• Import the workflow, assign credentials to all nodes, and replace any placeholders 🧩
• Run once to test end-to-end; adjust the 10s buffer if videos take longer ⏳