Convert Newsletters into AI Podcasts with GPT-4o Mini and ElevenLabs

🎧 Convert Unread Newsletters into Conversational AI Podcasts

Turn email overload into audio insights — automatically.

This workflow transforms unread newsletters sitting in your inbox into engaging, human-like audio conversations between two AI voices. It’s perfect for listening during your commute, workout, or while multitasking.

Inspired by Google's NotebookLM, this automation brings long-form content to life by summarizing dense text into a natural dialogue using OpenAI and generating high-quality voice narration with ElevenLabs. The result? A dynamic audio file sent right back to your inbox — hands-free, screen-free, and stress-free.

💡 What this workflow does

✅ Connects to your Gmail inbox to fetch unread newsletters 🤖 Uses GPT-4o Mini to summarize and rephrase content as a conversation 🗣️ Sends the dialogue to ElevenLabs to generate voice clips (voice1 + voice2) 🔁 Merges all audio segments into a single podcast-like MP3 using FFmpeg 📬 Emails the final audio back to you for easy listening

🛠️ What you'll need

A Gmail account with IMAP enabled An OpenAI API key (GPT-4o Mini recommended for cost/performance) An ElevenLabs API key + selected voice IDs A self-hosted or local n8n instance with FFmpeg installed Basic knowledge of binary data and audio handling in n8n

✨ Use cases

Convert long newsletters into hands-free listening experiences Repurpose Substack or Beehiiv content for podcast-like distribution Build an internal voice dashboard for teams who prefer audio updates

🙌 Want to go further?

This workflow is modular and extensible. You can add steps to:

Upload the final audio to Spotify, SoundCloud, or Telegram Publish to a private podcast RSS feed Create a daily audio digest from multiple newsletters

📬 Contact & Feedback

Need help customizing it? Have ideas or feedback?

Feel free to reach out:
📩 Luis.acosta@news2podcast.com

If you're building something more advanced with audio + AI, like automated podcast publishing to Spotify — let me know and I’ll figure out how I can help you!