Playground
Render a scene from one prompt.
Describe the scene like a director. Seed-Audio 1.0 performs, records and mixes it in a single pass, up to 2 minutes per take.
0 / 300Free web takes are capped at 300 characters.
Complete the check above to generate.
By generating you agree to the acceptable use policy. Output is AI-generated audio.
RENDERED TAKES
Real output from the exact prompts shown. Press play, then load one and make it yours.
Three generations, one breakfast table
Child, father and grandmother in a single take
00:18.9
Angry customer, calm support agent
Two voices, anger cooling into calm
00:39.2
Night-market argument, two vendors
Playful two-voice argument over market noise
00:14.5
Two-caster football finish
Two casters at full shout, stadium roar
00:13.2
Bedtime story, gentle elderly voice
Elderly whisper over a distant music box
00:15.5
Movie-trailer narrator
Single deep voice, cinematic hits
00:12.9
1940s radio-drama cold open
Narrator over vinyl crackle and jazz
00:19.5
Rain on a tin roof, solo cello
No speech, just rain, thunder and cello
00:31.5
Audiobook page, waves under the voice
Warm narrator, a chill on the last line
00:10.8
Two-host podcast cold open
Host banter over an intro sting
00:14.4
Fantasy-tavern NPC greeting
Dwarven innkeeper, fireplace and lute
00:13.9