Dialogue Scene
16:9A lawyer delivering a closing argument from the lectern of a wood-paneled courtroom, jury seated behind, synchronized speech
Seedance 1.5 Pro
ID:
bytedance/seedance-1.5-proSeedance 1.5 Pro video generation API by ByteDance — cinema-quality video with synchronized audio, multilingual dialogue, and up to 1080p.
Text to videoImage to videoAudio generationMultilingual1080p
Input
OutputView all
Output will appear here
480p
$0.26
720p
$0.52
1080p
$1.10
cURL
curl https://api.runbase.net/v1/runs \
-H "Authorization: Bearer $RUNBASE_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "bytedance/seedance-1.5-pro",
"input": {
"prompt": "A cinematic product photo of a ceramic lamp",
"aspect_ratio": "1:1",
"resolution": "1K"
}
}'Examples
Image to Video (Source)
16:9Extreme close-up of an older person's face by a window, a single tear forming, soft natural light
Image to Video (Result)
16:9The tear rolls slowly down the cheek, subtle facial movement, ambient room tone
Overview
Seedance 1.5 Pro is ByteDance's first video model with native audio generation, bridging the gap between the silent 1.0 series and the full-featured 2.0 release. It produces cinema-quality video at up to 1080p with synchronized dialogue, sound effects, and ambient audio in multiple languages. The model accepts up to two reference images for image-to-video, allowing first-frame and last-frame control. Durations are fixed at 4, 8, or 12 seconds.
Use cases
Dialogue-driven scenes — interviews, monologues, explainer videos with synced speech. Multilingual ad creatives where voiceover needs to match the visuals. Image-to-video with two reference frames to control both the start and end of a clip. Short narrative content with ambient sound design.
Inputs
All parameters are passed in the input object of the run request.
| Parameter | Required | Description |
|---|---|---|
| prompt | Yes | Text description (3–2500 chars) |
| aspect_ratio | No | Default 16:9. Options: 16:9, 9:16, 4:3, 3:4, 1:1, 21:9 |
| resolution | No | Default 720p. Options: 480p, 720p, 1080p |
| duration | No | Default 4. Options: 4s, 8s, 12s |
| generate_audio | No | Generate synchronized audio. Default false |
| image_urls | No | Up to 2 reference images (max 10 MB each) for image-to-video |
Prompt tips
Write dialogue as scene direction, not script
Rather than quoting speech ("He says: Hello"), describe the scenario: "A man greets someone warmly at a doorstep, casual tone." The model infers appropriate dialogue from context.
Use the two-image slot for start/end framing
Upload a first-frame image and a second image as the target end state. The model interpolates the motion between them, giving you tighter control over the arc of the clip.
Limitations
- Shorter prompt limit (2500 characters) compared to Seedance 2.0's 20000
- Only three fixed duration options (4s, 8s, 12s) — no arbitrary length
- Prompt minimum is 3 characters
- Audio generation increases per-run cost
- 1080p significantly increases generation time and cost compared to 720p
FAQ
What languages does the audio support?
The model generates multilingual dialogue and voiceover. Language is inferred from the prompt context — write your scene description in the target language or specify the language explicitly.
How is 1.5 Pro different from 2.0?
Seedance 2.0 supports arbitrary durations from 4 to 15 seconds, a 20000-character prompt limit, and generally higher visual fidelity. 1.5 Pro is limited to fixed 4/8/12s durations and 2500 characters. On Runbase, 1.5 Pro accepts up to two reference images (first and last frame), while 2.0 takes a single first-frame image.
Can I disable audio and just get video?
Yes. The generate_audio parameter defaults to false. Leave it off to get silent video output, same as the 1.0 models.