Characters sharing a scene
Write the lines, describe each voice and add the room around them. Generate a conversation with its background in one track.
AI AUDIO GENERATOR · SEED AUDIO 1.0
Describe a scene. Give it a sound.
Create character dialogue, narration, sound effects and background ambience in one audio track. Start with a description, or guide the result with an image or audio references.
Up to 120 seconds · Optional references · Audio and SRT download
Create your audio
Seed Audio 1.0Describe the sound, exact dialogue and timing. Up to 120 seconds per generation.0/3000
Audio settings
Download timed subtitles when the generated track contains speech.
Your audio
Hear your scene here
Your generated audio will appear here for playback, download and subtitle export.
Audio for the scene you are making
A sound inspired by your image
Upload a scene image and describe the words or sounds you need. Use a video keyframe as a visual reference for a new audio track.
Sound effects and atmosphere
Describe footsteps, weather, movement or a musical mood. For a full song with lyrics and structure, use the dedicated music generator.
From description to audio
Describe the scene
Write your script, setting, sounds and desired length.
Add an optional reference
Choose one image or up to three short audio references.
Generate and download
Listen to the track, download the audio and export available subtitles.
Questions about AI audio generation
What can I generate with Seed Audio 1.0?
You can request speech, character dialogue, sound effects, ambience and music in an audio track using a natural-language description.
Can I use an image as a reference?
Yes. Upload one JPG, PNG or WebP image up to 10 MB and describe the audio or spoken words you need. Image references cannot be combined with audio references or a preset voice.
Can I use a voice or audio reference?
You can upload up to three WAV, MP3 or OGG Opus audio references, each no longer than 30 seconds and no larger than 10 MB. Refer to them as @音频1, @音频2 and @音频3 in upload order. A preset voice, when available, occupies one slot.
How long can the generated audio be?
One generation can produce up to 120 seconds of original audio. Describe the total length and the timing of spoken lines in your prompt.
Can this page add synchronized audio to a video?
This page generates an audio file. You can use a video keyframe as an image reference, then combine the result with your video in an editor. Automatic video analysis, lip sync and video export are separate workflows.
Can I download subtitles?
Enable subtitles before generating. When timed speech subtitles are returned, you can download an SRT file alongside the audio.
Does it return separate dialogue, music and sound-effect tracks?
This integration returns one generated audio track. Generate elements separately if you want to mix dialogue, music and sound effects independently.
How are credits calculated?
The page shows the current rate before submission. Credits are reserved for up to 120 seconds, then settled against the original generated duration rounded up to a whole second. Unused credits are returned. Playback-speed changes do not change the original duration used for billing.