🎙️ Podcast Generator

If you enjoy this project and want to support my work, feel free to buy me a coffee ☕. Thank you for your support!

---

📖 Overview

A lightweight Python app with a modern UI that lets you generate multi-speaker podcasts from any script. It uses high-quality voices from the ElevenLabs and Google Gemini APIs, and can even create synchronized HTML demos of your podcasts.

🔑 Getting Started

Get your free API key from ElevenLabs (or Google Gemini). Add it securely to the app. Start creating your own podcasts in minutes.

💡 Note: I am an affiliate of ElevenLabs. Using this link may earn me a commission at no extra cost to you. ElevenLabs also offers a free tier with monthly character limits for premium voices.

See also the French README for a version in French.

---

✨ Features

• Modern UI: A clean, modern, and responsive interface built with customtkinter that adapts to your system's light or dark mode.
• Dual TTS Provider: Choose between the high-quality voices of Google Gemini or ElevenLabs.
• Synchronized HTML Demo: Automatically generate a shareable HTML page with your podcast audio and a synchronized, highlighted transcript.
• Flexible Formats: Export your creations in MP3 (default) or WAV formats.
• Customization: Configure and save voices for each speaker in your scripts, with options for language and accent.
• Voice Guides: Explore and listen to all available voices from Gemini and ElevenLabs directly within the settings. Add your favorite voices to your speaker list with a single click.
• Integrated Playback: Listen to and stop audio playback directly from the application (requires FFmpeg).
• Secure API Key Storage: Your Google Gemini API key is requested once and securely stored in your system's keychain (keyring).
• Accent and Language Support: Create podcasts in multiple languages with distinct voices and accents for each language (from the speaker settings with the ElevenLabs API or from the prompt with Gemini).

---

🌍 Multilingual Support

Thanks to the ElevenLabs or Google Gemini API, Podcast Generator supports multiple languages and accents, allowing you to:

• Create multilingual podcasts with distinct voices and accents for each language.
• Emotional tone adaptation from the script.
• Produce content for an international audience.
• Facilitate language learning with realistic dialogues.
• Enhance accessibility by tailoring language to your target audience.

---

💡 Use Cases

• Teaching and Training
  Transform your course materials or tutorials into multilingual audio podcasts to engage your learners.

• Content Creation
  Automate the production of podcasts in various languages to reach a broader audience.

• Accessibility
  Make your content accessible to a wider audience through multilingual support.

• Language Practice
  Create multilingual dialogues with distinct voices for each character.

---

💡 Examples

Creating multi-voice podcasts from written scripts

John: Hello everyone, welcome to this new episode.
Samantha: Today, we will explore the basics of artificial intelligence.
John: Stay with us to learn more!
Samantha: Don't forget to subscribe to our podcast.

⚠️ The speaker names in your script (e.g., John:, Samantha:) must exactly match the names defined in the "Voice Settings" window. If not, the script will not be processed.

Here's an example of a script with emotional tone instructions:

Read aloud the play with emphasing strongly the emotions of the characters.
Cyrano: [mocking] Ah ! non ! c’est un peu court, jeune homme !
On pouvait dire… oh ! Dieu ! … bien des choses en somme…
[sarcastic] Agressif : « moi, monsieur, si j’avais un tel nez,
Il faudrait sur-le-champ que je me l’amputasse ! »
[playful] Amical : « Mais il doit tremper dans votre tasse !
Pour boire, faites-vous fabriquer un hanap ! »
[dramatic] Descriptif : « C’est un roc ! … c’est un pic ! … c’est un cap !
Que dis-je, c’est un cap ? … C’est une péninsule ! »
[teasing] Curieux : « De quoi sert cette oblongue capsule ?
D’écritoire, monsieur, ou de boîte à ciseaux ? »

💡 Note on Annotations: The app uses square brackets [emotion] for ElevenLabs' emotional cues. If you use Gemini, the app will automatically convert them to parentheses (emotion) for you.

📦 Installation

1. Required Dependency: FFmpeg

For audio conversion and playback, this application requires FFmpeg to be installed on your system.

macOS

Install via Homebrew:

brew install ffmpeg

Linux

Most distributions provide FFmpeg in their package manager:

sudo apt install ffmpeg        # Debian/Ubuntu
sudo dnf install ffmpeg        # Fedora
sudo pacman -S ffmpeg          # Arch

Windows (Detailed Guide)

1. Download FFmpeg
   Go to the official FFmpeg build page:
   👉 https://www.gyan.dev/ffmpeg/builds/
   Download the latest "release full" ZIP archive (e.g., ffmpeg-release-full.7z or .zip).

2. Extract the archive

   • Right-click the downloaded ZIP file and choose Extract All…
   • You will get a folder named similar to ffmpeg-2025-xx-xx-full_build.

3. Move the folder

   • Move the extracted ffmpeg folder to a location where it will stay permanently, for example:
     C:\ffmpeg

4. Add FFmpeg to the PATH

   • Press Windows + R, type:

     sysdm.cpl
     

     and press Enter.
   • In the System Properties window, go to Advanced → Environment Variables.
   • Under System variables, find and select Path, then click Edit.
   • Click New and add the path to FFmpeg’s bin folder, e.g.:

     C:\ffmpeg\bin
     

   • Click OK to close all dialogs.

5. Verify installation

   • Open Command Prompt and type:

     ffmpeg -version
     

     You should see the version info, confirming FFmpeg is installed and accessible from anywhere.

---

2. Installing the Application

1. Go to the Releases page.
2. Download the latest archive for your system:
   • macOS/Linux: Download the .tar.gz file (..._MacOS_arm64.tar.gz for Apple Silicon, ..._MacOS_x86_64.tar.gz for Intel, or ..._Linux.tar.gz for Linux).
   • Windows: Download the .zip or .tar.gz file.
3. Extract the archive:
   • On macOS and Linux, double-click the .tar.gz file to extract the application, or use:

     tar -xzf Podcast_Generator_MacOS_arm64.tar.gz

   • On Windows, unzip the archive.
4. Move the application:
   • On macOS, drag Podcast Generator.app to your Applications folder.
   • On Windows, place the extracted folder anywhere you like.
   • On Linux, place it in your home directory or any preferred location.

---

💡 Note for macOS Users

When you first run the application, macOS will show several security warnings because it's not from the App Store. This is normal.

1. "App downloaded from the internet": Click Open.
2. "Unidentified Developer": macOS may block the app. Click OK, then:
   • Go to System Settings → Privacy & Security.
   • Scroll to the Security section.
   • Click Open Anyway and confirm.

---

💡 Note for Windows Users (Windows 10 / 11)

When you first run the application, Windows SmartScreen might block it because it’s not signed by Microsoft Store.

1. When you double-click the executable, you might see a window saying:
   "Windows protected your PC".

2. Click on More info.

3. Then click on Run anyway.

After doing this once, Windows will remember your choice and won’t show the warning again.

---

First Launch: API Key

On first launch, the application will request your ElevenLabs API key (and optionally your Google Gemini API key).
It will be stored securely.

---

🚀 Advanced Feature: HTML Demo Generation (optional)

The app can generate a shareable HTML page with your podcast audio and a synchronized, word-by-word highlighted transcript. This is perfect for sharing demos or for accessibility.

See an example of a generated HTML page here.

Installing Optional Dependencies for Demo Generation

This feature relies on whisperx for audio alignment, which requires PyTorch. These are heavy dependencies, so they are optional. To enable this feature, you need to install them manually.

1. Install PyTorch
It is highly recommended to install the CPU-only version of PyTorch, as it is much lighter and sufficient for this application. Visit the official PyTorch website and select the appropriate options for your system.

For example, using conda:

conda install pytorch torchaudio cpuonly -c pytorch

2. Install WhisperX
Once PyTorch is installed, you can install whisperx and its other dependencies using the [demo] extra: pip install .[demo]

pip install .[demo]

---

👨‍💻 For Developers

To contribute to the project, run the code, or create your own build, please refer to the full developer guide: ➡️ DEVELOPERS.md

---

📜 License

This project is licensed under the MIT License - see the LICENSE.md file for details.

---

🤝 Contributing

Contributions are welcome! Please see the DEVELOPERS.md file for guidelines.

---

Thank you for exploring Podcast Generator! Feel free to support this project with a coffee ☕.