Audio generation is primarily available on Pro+ plans. If audio generation is not accessible, upgrade in Settings → Account.
Step-by-step: generate audio
Describe the audio you want
Write a prompt describing your desired audio. Be specific about: genre, mood, tempo, instruments, duration, and how it will be used.
Select a model
Choose an audio generation model from the model dropdown. Different models handle music, voice synthesis, and sound effects differently — select the one that matches your use case. See audio models for guidance.
Set duration (if applicable)
If the selected model supports duration control, specify how long the audio should be. Include this in your prompt as well:
"Create a 45-second background track...".Click Generate
Click the Generate button. Audio generation typically takes a few seconds to about a minute depending on the model and duration requested.
Prompt examples by audio type
Background music for video
Background music works best when you describe the genre, mood, tempo, and intended emotional effect:Upbeat jazz piano loop for a 30-second product video, bright and energetic, 120 BPMCalm cinematic background track for a travel documentary, orchestral strings, 60 seconds, gentle swellUplifting corporate presentation music, modern, professional, subtle piano and light percussion, 90 secondsTense thriller-style underscore, low strings, minimal, building tension, 45 secondsWarm acoustic guitar loop, coffeeshop ambience, relaxed and friendly, 60 seconds
Ambient soundscapes
Calm ambient soundscape with rain and distant thunder, indoor atmosphere, no musicForest environment: birds chirping, wind through leaves, distant stream, peaceful morningBusy city street ambience: traffic, distant conversations, city energy, urban daytimeUnderwater ambience: gentle bubbling, muffled resonance, serene and spacious
Intros and jingles
Short 5-second intro jingle for a tech podcast, modern, professional, upbeat10-second outro music with fade, light and friendly, optimistic3-second notification sound: positive, clear, modern UI styleBrand jingle: catchy, memorable, 8 seconds, fun and approachable
AI narration / text-to-speech
Read the following text in a warm, professional male voice at a medium pace: [text]Narrate in an enthusiastic, energetic tone suitable for a product demo: [text]Calm, soothing female voice for a meditation guide introduction: [text]News anchor style, clear and authoritative: [text]
Sound effects
Single camera shutter click, professional DSLRTyping on a mechanical keyboard, brief burst, 2 secondsCoin drop on hard floor, metallic ringPositive success chime, UI sound, cheerfulDoor opening and closing, wooden door, interior
Output formats
Audio is available for download in standard formats depending on the selected model:| Format | Best for |
|---|---|
| MP3 | Universal sharing, web, social media, mobile |
| WAV | Professional audio workflows, lossless quality, video editing |
| Other formats | Model-dependent — check the download options |
Tips for better audio
- Specify duration explicitly:
"Create a 60-second..."or"a short 5-second..."helps the model target the right length - Describe the intended use: telling the model what the audio is for (“background music for a product video”, “notification sound for a mobile app”) helps it understand context and emotional register
- Name instruments specifically: “piano and strings” is better than “music”; “acoustic guitar” is better than “guitar”
- Describe tempo in BPM or adjective:
"120 BPM","slow and deliberate","energetic and fast-paced" - Mention what to avoid:
"no lyrics","no drums","instrumental only"helps prevent unwanted elements

