Back to Templates

Create LinkedIn content from Telegram voice/text using Whisper, GPT-4 & Dumpling AI

Created by

Created by: Yang || yang
Yang

Last update

Last update 3 months ago

Share


🧾 What this workflow does

This automation turns a Telegram voice or text message into a high-quality LinkedIn post using OpenAI’s Whisper, Dumpling AI and GPT-4, and then repurposes that same message into a content idea with visual inspiration using Dumpling AI. All outputs are saved to Airtable for review, approval, and scheduling.


šŸ‘¤ Who is this for

  • Founders or creators sharing ideas while walking or multitasking
  • Content teams turning voice memos into carousel content
  • Personal brands posting on LinkedIn with minimal manual work
  • Virtual assistants or SMMs handling voice-driven content tasks

āš™ļø How to set up

āœ… Requirements

You’ll need active accounts for:

  • Telegram Bot (via n8n’s Telegram Trigger node)
  • OpenAI (for Whisper and GPT-4)
  • Dumpling AI (to search Google and create Image)
  • Airtable (to log your posts)

In Airtable, create a table with the following fields:

  • Ideas (Single line text)
  • Context (Long text)
  • Image URL (URL)

šŸ”§ Setup steps

  1. Set up your Telegram bot

    • Use BotFather to create a bot and get your token
    • Add the bot to a private group or channel
  2. Connect all credentials in n8n

    • Telegram Trigger (Bot Token)
    • OpenAI (for Whisper + GPT-4)
    • Dumpling AI (HTTP Auth)
    • Airtable (OAuth or Personal Token)
  3. Update node parameters

    • In the Telegram Trigger node, specify your bot and listening chat
    • In the HTTP node for Dumpling AI, add your API key via header
    • In Airtable nodes, map to your base and table correctly
  4. Optional prompt customization

    • In the GPT-4 node, you can update the prompt for tone, formatting, or CTA style
  5. Optional: Adjust the Dumpling AI prompt to suit your niche (e.g., health, tech, finance).


🧠 How it works

  1. The Telegram bot receives a voice message.
  2. Whisper transcribes the voice into clean text.
  3. GPT-4 formats the text into a polished LinkedIn post.
  4. Dumpling AI:
    • Generates a short-form content idea from the transcript.
    • Creates a branded image prompt and renders a matching image.
  5. The full set of outputs (text, idea, image) is saved to Airtable.
  6. You can now review and schedule the post manually or via another automation.

šŸ› ļø Customization ideas

  • Swap Telegram for WhatsApp or voice memos via Google Drive
  • Add moderation or approval steps before saving to Airtable
  • Send the LinkedIn post directly to Buffer or Hootsuite for auto-posting
  • Generate post captions or hashtags using GPT-4 before publishing
  • Extend the workflow to auto-create carousel slides from the content idea

This workflow makes it easy to share your voice with your audience — literally. Whether you're walking, thinking out loud, or sending team updates, this automation turns your spoken ideas into professional content, ready to post or polish.