Stable Audio
Stable Audio is Stability AI's music generation model, producing full stereo tracks and sound effects from text prompts. It grew out of the same research culture as Stable Diffusion, and an open version of the model is available for non-commercial experimentation.
Quick Facts
| Developer | Stability AI |
|---|---|
| First released | 2023 |
| Latest version | Stable Audio 2.0 (2024) |
| Platform | Web browser |
| Output | Full stereo tracks and sound effects |
| License | Proprietary; open model weights for non-commercial use |
| Pricing | Freemium; exact plans on the official website |
What is Stable Audio?
Stable Audio is a generative music tool from Stability AI, the company behind the famous Stable Diffusion image models. You describe the track you want and it produces finished audio, including full songs and individual sound effects, in a web-based editor.
The first version arrived in September 2023, and Stable Audio 2.0 followed in 2024 with longer generations, full stereo output and higher quality. Stability also released Stable Audio Open, a version of the model with open weights, aimed at researchers and developers who want to fine-tune or study the technology.
Why it matters: Stability's approach has always been about open research pushing the field forward. Stable Audio brings that spirit to music, giving creators a serious alternative to traditional sample libraries.
Key Features
- Text to music generation for full stereo tracks
- Sound effects generation for game and video production
- Longer generations in the 2.0 model, up to several minutes
- Web-based editor with prompt-based control
- Open model weights for non-commercial research use
- Integration with Stability's broader AI platform
How to get started
The easiest path is the web app at stableaudio.com, where a free account gives you a monthly allowance of generations. Type a detailed prompt describing genre, mood and instruments, then generate and download.
- Create a free account at stableaudio.com
- Choose music or sound effects mode
- Write a detailed prompt with style and mood
- Generate and preview the results
- Download the audio or export it into your project
Use cases
- Background music for videos, streams and presentations
- Sound effects and ambience for game development
- Quick music beds for podcasts and ads
- Research into generative audio with the open model
- Prototyping sound design before recording the real thing
Pricing and licensing
Stable Audio uses a freemium model: a free tier with monthly credits, and paid memberships for more generations and commercial usage. Exact plan prices live on the official website.
The open Stable Audio Open model has a separate non-commercial licence, while the web service requires a paid plan for commercial work. Read the licence that matches the version you use before shipping a project.
Pros and cons
The audio quality is strong for a generative tool, and having both music and sound effects in one service is handy. The open model is a genuine gift to researchers and tinkerers.
Control is still prompt-based, so you cannot fine-tune arrangements the way you would in a DAW, and the free tier is quite limited. Stability's frequent product changes also mean features and pricing shift regularly.
Alternatives
- Suno: full songs with vocals from a text prompt
- Udio: music generation with fine control over structure
- AIVA: composed soundtracks for games and film
- Soundraw: royalty-free tracks you generate and customise
- LANDR: AI mastering and distribution for finished music