A modern, React-based frontend application for generating podcasts using AI text-to-speech models.
- Podcast Generator: Create podcasts with up to 4 speakers
- Voice Library: Upload, manage, and delete custom voice files
- Multiple AI Models: Support for Chatterbox Pro, Vibe Voice, and FishAudio
- Dynamic Settings: Model-specific parameters that change based on selected AI model
- History: Track all generated podcasts with metadata
- Modern UI/UX: Clean design with Talkai branding and color scheme
ttsui/
├── public/
│ └── index.html
├── src/
│ ├── components/
│ │ ├── Sidebar.js
│ │ ├── Sidebar.css
│ │ ├── Settings.js
│ │ └── Settings.css
│ ├── pages/
│ │ ├── PodcastGenerator.js
│ │ ├── PodcastGenerator.css
│ │ ├── VoiceLibrary.js
│ │ ├── VoiceLibrary.css
│ │ ├── History.js
│ │ └── History.css
│ ├── styles/
│ │ └── App.css
│ ├── App.js
│ └── index.js
└── package.json
- Navigate to the project directory:
cd ttsui- Install dependencies:
npm installStart the development server:
npm startThe application will open at http://localhost:3000
Create an optimized production build:
npm run build- Exaggeration: Control voice style exaggeration (0-1)
- CFG Weight: Classifier-free guidance weight (0-10)
- Temperature: Randomness in generation (0-2)
- CFG Weight: Classifier-free guidance weight (0-10)
- Speed: Playback speed (0.5-2)
- Top K: Top-k sampling parameter (1-100)
- Temperature: Randomness in generation (0-2)
- Repetition Penalty: Penalty for repetition (0-2)
- Primary: Dark Navy (#1e3a5f, #2c4d7a)
- Accent: Teal (#00b8a9)
- Background: Light Gray (#f5f5f5)
- Text: Dark (#1a1a1a)
- React 18.2.0
- React Router DOM 6.20.0
- CSS3 with modern features
- No additional UI libraries for cleaner, custom design
- Chrome (latest)
- Firefox (latest)
- Safari (latest)
- Edge (latest)