Stable Audio 3 for Text to Music

ABOUT THE WORKFLOW

Make Audio From Text
Write a short description of the music or sound you want, set how long it should be, and the workflow returns a finished audio track. No instruments, no recording gear, no editing software. A newcomer can get a usable result on the first run.

Model

• Stable Audio 3 Medium by Stability AI. An open-weight text-to-audio model that generates instrumental music and sound effects from a written prompt, at lengths up to about six minutes.

HOW IT WORKS

Step 1. Write your prompt
Describe the audio you want. A short phrase is enough to start, and the workflow expands it into a fuller prompt for you.
Works great with: music · sound effects · ambient beds

Step 2. Set the type and length (optional)
Choose what you are making (Music, Instrument, SFX, or One-shot) and how many seconds it should run. Leave both on their defaults for a standard music track.

Step 3. Hit run and download
The model generates the audio and returns an MP3. Preview it in the workflow, then download.
Ready for: video editors · game engines · podcast tools · any DAW

First time? Leave every setting as-is. The defaults (Music · 150 seconds · prompt enhancer on · random seed) are the right starting point for almost everyone.

RECOMMENDED SETTINGS

Quick-start guide. Find the goal that matches yours and copy the settings.

• Standard music track (most people) — Music type · 150 seconds · prompt enhancer on · random seed. The right starting point for almost everyone.

• Want a fast test first — Drop the length to 20 to 30 seconds. You hear the direction of a result in less time before committing to a full-length run.

• Making a sound effect, not music — Switch the type to SFX. Use One-shot for a single hit or a short sample rather than a full arrangement.

• Want the exact same track again — Set a specific seed instead of leaving it random. The same prompt and seed reproduce the same audio.

• Want the model to follow your words exactly — Turn the prompt enhancer off. Your description goes in unchanged instead of being rewritten into a fuller prompt.

• The result is not matching your prompt — Name the genre, instruments, mood, and tempo in BPM. Thin prompts produce thin audio.

Prompt: Describe genre, instrumentation, mood, and tempo, not only the subject. "Warm lo-fi hip hop with a dusty piano loop, soft vinyl crackle, and a mellow head-nod groove at 85 BPM" gives the model far more to work with than "chill beat." Adding a BPM and one or two named instruments is the fastest way to sharpen a result.

LEARN

📹 Videos

• Intro to Floyo

• ComfyUI 101 Free Course ft. Sebastian Kamph

• Floyo 101 for Team Collaboration

✨ Quick links

• Help docs

• LoRA Training guide

• Discover more workflows

USE CASES

🎬 Video Editors & Filmmakers
Generate a soundtrack bed that fits the mood of a scene, then drop it straight onto your timeline. No licensing search, no royalty tracking.

🎮 Game Developers
Produce background loops, ambient textures, and sound effects for prototyping and level design in seconds, without a composer on call.

🎙️ Podcasters & Creators
Make custom intros, outros, and transition stingers that match your show. Set a short length and generate a few options to pick from.

🎵 Musicians & Producers
Sketch instrumental ideas, loops, and genre studies fast. Use them as a starting point, a reference, or a bed to write over.

WHAT WORKS BEST / WHAT TO AVOID

✅ Works great

• Prompts that name genre, instruments, mood, and BPM

• Instrumental music and sound effects

• Ambient beds, loops, and background tracks

• Short test runs before a full-length generation

⚠️ May produce softer results

• Vague prompts like "make something good"

• Requests for vocals, lyrics, or spoken word

• Full radio-ready songs with a lead singer

• Prompts with several conflicting genres at once

FAQ

What is Stable Audio 3 Medium?
Stable Audio 3 Medium is an open-weight text-to-audio model from Stability AI, released in 2026. It generates instrumental music and sound effects from a written prompt using a latent diffusion architecture, and it can produce tracks up to about six minutes long. It is the open Medium variant of the Stable Audio 3 family, trained on fully licensed audio data.

Can Stable Audio 3 generate vocals or full songs with lyrics?
No. Stable Audio 3 is built for instrumental music and sound effects, not for vocals, lyrics, or speech synthesis. It handles arrangements, ambient textures, and genre-driven backing tracks well. If you need sung vocals and lyrics, tools like Suno or Udio are aimed at that instead.

How long can the generated audio be?
Up to about six minutes. The model sizes each generation to the length you request, so you can make a two-second sound effect or a multi-minute track from the same workflow. This one defaults to 150 seconds, and you can raise or lower it.

Can I use Stable Audio 3 music commercially?
Yes, under Stability AI's Community License you own and can commercialize the outputs you generate. The models are trained on fully licensed data (licensed catalog audio plus Creative Commons recordings), which is why commercial use is covered. Organizations above $1M in annual revenue should review Stability AI's Enterprise license for their terms.

How is Stable Audio 3 different from Suno or Udio?
Suno and Udio are built to produce full songs with vocals and lyrics from a prompt. Stable Audio 3 focuses on instrumental music and sound effects, and it ships as open weights so it can run in a ComfyUI workflow like this one. If your project needs a backing track, ambient bed, or SFX rather than a sung song, this is the closer fit.

What makes a good Stable Audio 3 prompt?
Name the genre, the instruments, the mood, and the tempo in BPM. "Driving synthwave with analog bass, gated reverb drums, and a bright arpeggio at 110 BPM" gives the model far more to shape than "retro music." Adding a BPM and one or two specific instruments is the single biggest lever on quality.

How to run Stable Audio 3 online?
You can run Stable Audio 3 online through Floyo. No installation, no setup, no API key to wire up. Open the workflow in your browser, type a prompt, and hit run. Free to try.

WHY FLOYO?

Floyo is the only platform with team collaboration for ComfyUI in the browser. You run workflows with no install. You share run history, assets, and models across your team. You pay only when you generate. Floyo supports open-source and closed-source models.

A sound designer generates a track and likes the result. A teammate opens that exact run from shared history and keeps going. No file handoffs. No version confusion.

For studios and enterprise teams, Floyo adds private workspaces, pooled resources, and a team usage dashboard. Other ComfyUI cloud tools run for one person at a time. Floyo runs for the whole team, with transparent per-generation costs.

Ready to try it?
Type a prompt, set a length, and run it. The settings are already dialled in.

→ Launch Workflow, Free

Questions? Watch the free course or check the FAQ above.