Piper TTS Dashboard 🎙️

A premium, web-based dashboard for the Piper neural text-to-speech engine. This application provides a high-quality interface for generating speech, managing voice catalogs, and organizing audio output.

🚀 Key Features

• MP3 Output Support: Automatically converts WAV output to high-quality MP3 (192kbps) for better compatibility and smaller file sizes.
• Hindi Language Support: Integrated 3 Hindi voices (hi_IN) by updating to the latest Piper catalog.
• 0-Byte Bug Fix: Implemented in-memory buffering and validation to ensure no empty audio files are generated.
• 100+ Voices: Access to the full Piper voice library, filterable by language and quality.
• Glassmorphism UI: A premium dark-mode interface with smooth animations and responsive design.
• Background Synthesis: Real-time progress tracking for long scripts.
• File Management: Integrated library to play, download, and delete generated files.
• Advanced Controls: Fine-tune speech with Length Scale, Noise Scale, and Sentence Silence parameters.

🛠️ Technology Stack

• Backend: FastAPI (Python), piper-tts 1.2.0, pydub, ffmpeg.
• Frontend: Vanilla HTML5, CSS3 (Modern Glassmorphism), JavaScript (ES6+).
• Core Engine: Piper (ONNX-based neural TTS).

📋 Setup & Installation

Prerequisites

• Python 3.9+
• FFmpeg (Required for MP3 conversion)

Steps

1. Clone & Navigate:

   cd piper/dashboard

2. Install Dependencies:

   pip install -r requirements.txt

3. Run the Dashboard:

   python app.py

4. Access the UI: Open http://localhost:8787 in your browser.

📂 Project Structure

• app.py: Main FastAPI backend handling API logic and synthesis tasks.
• static/: Frontend assets (HTML, CSS, JS).
• models/: Automatically downloaded voice models (.onnx and .json).
• output/: Directory for generated MP3 files.
• voices.json: The latest Piper voice catalog fetched from HuggingFace.

⚖️ License

This project inherits the license of the Piper TTS engine (MIT).