audiocraft-audio-generation

Generate music and sound effects from text prompts using AudioCraft models.

Updated Apr 9, 2026

Search Tags:#audio-generation #text-to-sound #text-to-music #audiocraft #musicgen #audiogen #encodec

Ready to run

npx skills add https://github.com/MarbleSodas/Mavis --skill audiocraft-audio-generation-marblesodas

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: audiocraft-audio-generation
Source: https://github.com/MarbleSodas/Mavis/tree/main/skills/mlops/models/audiocraft
Command: npx skills add https://github.com/MarbleSodas/Mavis --skill audiocraft-audio-generation-marblesodas

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

AudioCraft enables fast, high-quality text-to-audio and text-to-sound generation so you can turn creative prompts into playable audio without manual studio production.

Core Features & Use Cases

Text-to-Music (MusicGen): Generate music from detailed prompts, including melody-conditioned and stereo variants.
Text-to-Sound Effects (AudioGen): Create short sound effects like ambience, impacts, and environmental audio from text.
Neural Audio Codec (EnCodec): Support high-fidelity audio encoding/decoding workflows and deeper experimentation.
Use Case: A content team needs 10 distinct background tracks and 20 environment SFX variants for a short video; generate them directly from prompt text and save to WAV for editing.

Quick Start

Use the audiocraft-audio-generation skill to generate a 10-second track from the prompt "upbeat electronic dance music with punchy drums at 128 bpm" using the MusicGen text-to-music workflow.

Frequently Asked Questions about audiocraft-audio-generation

High-intent search queries and answers about installing and using this skill.

FAQPage Schema

How do I generate music from text prompts using AudioCraft?▼

To generate music from text prompts using AudioCraft, you use the MusicGen model to process natural-language descriptions and output playable WAV audio. You configure generation parameters like duration, select a pretrained model variant, and save the resulting audio at the correct sample rate.

Can I create sound effects from text with AudioGen?▼

Yes, AudioGen creates short sound effects from text. It is designed to generate environmental audio, ambience, and impacts, allowing you to produce SFX variants directly from prompt text for creative workflows without manual studio production.

What do I need to install to use AudioCraft for text-to-audio generation?▼

You need to install the audiocraft package along with PyTorch and transformers. This environment setup allows you to load pretrained model variants and execute text-to-music or text-to-sound generation workflows.

Does AudioCraft support melody-conditioned and stereo music generation?▼

Yes, AudioCraft supports melody-conditioned and stereo music generation through MusicGen variants. You can provide an existing melody to guide the output structure and generate high-fidelity stereo audio tracks from your text prompts.

What is the difference between MusicGen and AudioGen in AudioCraft?▼

MusicGen is used for text-to-music generation, creating full tracks from prompts, while AudioGen is used for text-to-sound effects, creating short environmental audio and impacts. Both output WAV files but serve distinct creative production needs.

Are there limitations when generating audio with AudioCraft models?▼

AudioCraft generation requires configuring correct sample rates to save WAV files properly and relies on pretrained model variants. It is built for short sound effects and music generation, meaning output length is constrained by your configured generation parameters.