Stable Audio

Overview

Stable Audio is an audio generation model built by Stability AI that produces music and sound from written Prompts or from audio references you supply. Rather than stitching together Samples, it synthesizes finished stereo audio, which makes it useful for quick sketching, soundtracking short vIDEos, and generating ideas that can be refined later in a DAW.

What it does

You describe the sound you want in plAIn language, for example a genre, mood, tempo, or instrumentation, and the model returns a complete audio CLIp. The SECond generation of the model extended the usable clip length and raised ouTPUt quality, with generation reported at 44.1 kHz stereo. It can also take an existing audio input and build on it, which supports variations and extensions of an idea.

Key capabilities

• text-to-music: Prompt by genre, mood, tempo, and instruments to get a rendered track.
• Audio-to-audio: Feed in a reference clip and generate material in a related style.
• Broad genre coveRAGe: Supports many styles, including rock, jazz, electronic, hip-hop, folk, pop, and country, with more than twenty genres commonly cited.
• Longer clips: Output lengths are measured in minutes rather than seconds, making clips more usable for real content.
• sound effects: The model is not limited to music and can render ambient and effect-style audio.
• Commercial-friendly terms: Stability AI has stated that generated audio may be used commercially, though the exACT terms should be checked against the current license.

Who it is for

Music producers use it for fast ideation and for generating loops or beds to build on. Video creators and podcasters use it to add background music without licensing headaches. Advertisers and marketers use it to draft tracks that fit a brand mood. Game and APP developers use it for placeholder or production audio. Hobbyists use it to explore styles they cannot eASIly perform themselves.

How to get started

1. Open the Stable Audio web app, or reach the model through Stability AI's API and platform partners.
2. Write a prompt that names genre, instrumentation, tempo, and mood, and optionally set a duration.
3. Generate, audition, and download the resulting clip.
4. Refine with follow-up prompts or by feeding your favorite result back in as audio input.
5. Use a DAW for editing, mixing, and mastering if the track is headed for release.

Pricing and access caveat

Access options and pricing have changed over time and may vary by region and platform, so confirm the current plans, credit allowances, and commercial license terms before putting generated audio into a paid project. Voice and music models also raise rights questions, so review Stability AI's usage policy alongside your own distribution requirements.