AudioCraft

What is AudioCraft?

AudioCraft is a code base from Meta AI research for generative audio: music, sound effects and compression, all trained on raw audio signals. MusicGen produces music samples from text prompts, and AudioGen produces environmental sounds from text. Both are single autoregressive language models that work over streams of discrete audio tokens, and a token interleaving pattern lets one model capture long-range structure in the audio. The tokens come from EnCodec, a neural audio codec that compresses audio into parallel streams, and its decoder turns generated tokens back into a waveform. EnCodec supports 48 kHz stereo and 24 kHz mono audio, and a Multi-Band Diffusion option reduces artifacts at the cost of more computation. The demo site has audio samples and links to the code, and training code for EnCodec has been released. It is a toolkit for developers and researchers, not a point-and-click music app, and the site shows no pricing.

Use cases

• Generate music from a text prompt
• Create environmental sound effects from text
• Compress audio with a neural codec
• Study generative audio model design
• Train a custom audio compression model

Who it's for

• AI researchers
• Machine learning engineers
• Audio engineers
• Music technologists

Why use AudioCraft

• Get music, sound-effect and codec components in one code base instead of separate projects.
• Study a simple design: one language model over compressed tokens handles both music and sound.
• Listen to demo samples on the site before setting anything up.
• Train your own EnCodec compression model, because the training code has been released.
• Pick codec bitrates from 1.5 to 24 kbps to balance size against fidelity.

AudioCraft features

• MusicGen for text-to-music generation
• AudioGen for text-to-sound generation
• EnCodec neural audio codec producing parallel streams of discrete tokens
• Single autoregressive language model over audio tokens with a token interleaving pattern
• 48 kHz stereo EnCodec model at 3, 6, 12 and 24 kbps
• 24 kHz mono EnCodec model at 1.5, 3, 6, 12 and 24 kbps
• Multi-Band Diffusion decoder for fewer artifacts at low bitrates
• Released EnCodec training code and audio demo samples

How to get started with AudioCraft

1. Listen to the MusicGen, AudioGen and EnCodec samples on the AudioCraft demo site.
2. Follow the link to the code and set up a Python environment with the dependencies.
3. Load MusicGen or AudioGen and supply a text description as the conditioning input.
4. The language model generates audio tokens, and the EnCodec decoder converts them back into a waveform.
5. Save the audio, or train your own EnCodec model for compression with the released training code.

AudioCraft limitations

• A research code base that needs coding skills and compute, not a consumer app
• No pricing or licensing terms appear on the demo site, so check the code repository before commercial use
• Multi-Band Diffusion lowers artifacts but needs more computation

AudioCraft pricing

We couldn't confirm pricing on the official site. Check audiocraft.metademolab.com for current prices.

Prices can change and may vary by country or exclude tax. Check the official pricing page.

AudioCraft alternatives

The main AudioCraft alternatives listed on Debuly are Stable Audio, Suno, Mubert, Beatoven.ai, Hume AI and Loudly. The table compares pricing and platforms; order is A–Z.

1. 

   Beatoven.ai

   AI music generator for royalty-free background tracks and text-to-sound-effects, with MP3 or WAV downloads and a monetization license.

   Audio and voiceVideo
   PaidVisit Beatoven.ai (official site)
2. 

   Hume AI

   Expressive text-to-speech and empathic voice agent APIs that let developers direct emotion, tone and pacing in generated speech.

   Audio and voiceDeveloper platformsAutomation and agents
   FreemiumVisit Hume AI (official site)
3. 

   Loudly

   AI music platform with a music generator, text-to-music, remixer, stem splitter, distribution to streaming services and a developer API.

   Audio and voiceVideoDeveloper platforms
   Visit Loudly (official site)
4. 

   Mubert

   AI music generator for royalty-free tracks and streams, with a web app for creators and an API for developers.

   Audio and voiceDeveloper platformsVideo
   FreemiumVisit Mubert (official site)
5. 

   Stable Audio

   Stability AI's music and sound generator with a multi-track deck, audio-to-audio editing, a DAW plugin and an API, billed in credits.

   Audio and voiceDeveloper platforms
   Visit Stable Audio (official site)
6. 

   Suno

   AI music generator that creates full songs with vocals, lyrics and production from a text prompt, on web and mobile.

   Audio and voice
   FreemiumVisit Suno (official site)

AudioCraft FAQ

Sources & verification

Facts on this page come from the official sources below, each with the date we read it. Written by Debuly; the vendor did not review this page.

Image credits

• Logo: from audiocraft.metademolab.com · fetched 23 Oct 2026 · Vendor trademark, used only to identify the product; removed on request
• Cover: none collected yet, so a generated tile is shown.

Spot an error? Report an issue with this listing. We also remove a logo or image on the owner's request.