UNLIMITED Seedream 5.5 · Live on every paid plan

See how
Cover for the guide on making an AI video with sound
Audio

Official BytePlus partnerBytePlus, a ByteDance company

How to make an AI video with sound

Make an AI video with sound: write ambience and dialogue into the prompt, add audio references, or lay music in the editor. Generated audio adds no credits.

KreatorFlow3 min read

Short answer

Seedance clips come with sound. Write the sound into the prompt in the same way as the picture: what you hear, how loud and when. Add up to 3 audio references if you want a specific voice or track. Generated audio adds no credits, so a 5-second clip is 26 credits on Seedance 2.5 at 720p.

Updated:

Add sound in four ways

  1. 1

    Describe the ambience

    End the prompt with a sound line: rain on a window, a busy market, quiet wind. The model makes the sound with the picture, so it fits the scene.

  2. 2

    Write short speech in quotes

    For a spoken line, put it in quotes and say who speaks: a woman says, "We made it." Keep it to one or two short sentences per clip.

  3. 3

    Use an audio reference

    Add a voice or a track as an audio reference. The studio takes up to 3 references, each at least 2 seconds, and Seedance 2.5 accepts up to 30 seconds of audio.

  4. 4

    Finish in the editor

    Add a music track, a voiceover or captions on the timeline, then export. This is the way to get a long, steady track under many clips.

Prompts with sound

The last sentence of each prompt is the sound.

  • Ambience

    A small fishing boat rocks at dawn in a quiet harbour. Mist on the water. The camera drifts slowly. Sound of gentle waves, creaking wood and distant gulls.

  • Speech

    A chef in a small kitchen looks at the camera and says, "Taste this." He lifts a spoon toward the lens. The camera is steady. Sound of a simmering pot.

  • Music beat

    A dancer in an empty studio moves in time to a slow, deep beat. The camera circles her once. Soft drum beat and a low synth.

Sound made with the picture

Because the sound is made together with the clip, a door that closes makes a door sound at that moment. You do not have to sync it by hand.

Be specific. Rain is vague. Heavy rain on a tin roof, with distant thunder, is a clear sound for the model to follow.

When to use a separate voice

For narration, a long speech or a language you need to control, make the voice in the voice tool and put it on the timeline. The sound inside the clip then stays as quiet ambience.

To make a face speak your own recording, use the lip sync tool.

Music

Use music you own or have the right to use. Add it in the editor, trim it to the cut, and lower it under speech.

Credits

A 5-second clip at 720p is 26 credits on Seedance 2.5 and 17 on Seedance 2.5 Flash, with sound. Sound does not add to the price.

Frequently asked questions

How do I make an AI video with sound?

Seedance clips come with sound. Write the sound into the prompt in the same way as the picture: what you hear, how loud and when. Add up to 3 audio references if you want a specific voice or track. Generated audio adds no credits, so a 5-second clip is 26 credits on Seedance 2.5 at 720p.

Does the video come with sound by default?

Yes. The clip is made with sound, and you can steer it with a sound line in the prompt.

Does sound cost extra credits?

No. Generated audio adds no credits on Seedance 2.x.

Can I add my own audio?

Yes. Add up to 3 audio references in the studio, or lay a track on the timeline in the editor.

Can the people in the clip speak?

Yes. Write the line in quotes and say who speaks. Keep it short.

How do I add a voiceover?

Make the voice in the voice tool and add it on the timeline in the editor.

Make a clip with sound

Write the sound into the prompt and see the credits first.

More guides