Back to catalog
MusicGen (Meta)
Meta transformer-based open-source model generating high-quality instrumental music snippets.
This profile summarizes information listed in the Zancle catalog. Pricing, features, formats, APIs, and licensing terms can change: confirm decision-critical details on the official website before purchasing or using the tool commercially.
Overview
MusicGen, developed by Meta AI Research (FAIR), is a state-of-the-art deep learning model capable of generating stereo music from text prompts or reference melodies. Trained on thousands of hours of licensed music, it can run locally for developers and producers.
Listed features
Text-prompted or melody-guided audio generation
Multiple model sizes (Small, Medium, Large)
High-quality signal processing via EnCodec
Free local execution or cloud demo spaces (Hugging Face)
Points to consider
100% free and open-source
Ability to guide generation with a hummed or whistled melody
Excellent acoustic instrument fidelity
Listed limitations or constraints
Requires a powerful GPU to run locally
Generation duration limited by default in online demo
Listed formats
License information listed in catalog
Licence de recherche et d'utilisation commerciale autorisée (selon modèles de Meta)
Check official terms