Text-to-Music Generation App

Text-to-Music Generation App

Category: Other
License: MIT
Model Type: Speech Synthesis
A Streamlit-based web application built on Meta’s MusicGen (part of the Audiocraft library) that converts natural language prompts into short musical compositions. The app offers an intuitive interface to explore AI-generated music using high-quality pretrained models in a lightweight Python environment.

Key Features

  • Generates music from text prompts using MusicGen Small model
  • Adjustable duration slider (2–20 seconds)
  • Live audio playback in the browser
  • Downloadable WAV files of generated music
  • Automatic model loading and caching for speed
  • Minimal dependencies using Streamlit, Audiocraft, and Torchaudio
  • Single-script deployment with simple setup

Project Screenshots

Project Screenshot
Project Screenshot