Audio
Meta MusicGen
Use it ↗Meta MusicGen generates music from text and can be guided by a reference melody, creating up to 30-second clips. Pricing starts from 15 CUs.
Meta MusicGen is a legacy 2024 text-to-music model with a unique dual-input capability, generating music from text alone or a combination of text and a reference audio melody. Its primary strength is this “melody conditioning” feature, which lets creators reimagine existing melodies in new styles or maintain thematic consistency. This makes it effective for musical exploration and production, though its generation length is limited to 30 seconds. The model’s architecture is built on the open-source AudioCraft library.