ComfyUI Node
Seedance – Video Generator
A ComfyUI node in Seedance/Video Generation with 27 inputs and 6 outputs.
Seedance – Video Generator
- first_frame
- last_frame
- reference_image_1
- reference_image_2
- reference_image_3
- reference_image_4
- reference_video
- reference_audio
- video
- last_frame
- frames
- video_url
- video_path
- last_frame_url
◄api_key►
◄promptA girl holding a fox, the camera slowly pulls out.►
◄modeldreamina-seedance-2-0-260128►
◄ratio16:9►
◄duration5►
◄resolutiondefault►
◄generate_audiofalse►
◄watermarkfalse►
◄return_last_frametrue►
◄poll_interval10►
◄max_wait600►
◄first_frame_url►
◄last_frame_url►
◄ref_url_1►
◄ref_url_2►
◄ref_url_3►
◄ref_url_4►
◄reference_video_url►
◄reference_audio_url►
CategorySeedance/Video Generation
Inputs (27)
| Name | Type | Default | Description |
|---|---|---|---|
| api_key | STRING | BytePlus ARK API key. Leave empty to use ARK_API_KEY env variable. | |
| prompt | STRING | A girl holding a fox, the camera slowly pulls out. | Text description of the video. Use [Image 1-4] for reference images, [Video 1] for reference video, [Audio 1] for reference audio. |
| model | COMBO | dreamina-seedance-2-0-260128 | dreamina-seedance-2-0-260128 → T2V / first-frame / first+last modes. seedance-1-0-lite-i2v-250428 → reference image mode. |
| ratio | COMBO | 16:9 | 6 options: 16:9, 9:16, 1:1, 4:3, 3:4, adaptive |
| duration | INT | 54–15 | Video duration in seconds. Seedance 2.0 supports 4–15 s. seedance-1-5-pro-251215 is capped at 12 s. |
| resolution | COMBO | default | Output resolution. 'default' lets the API decide. |
| generate_audio | BOOLEAN | false | — |
| watermark | BOOLEAN | false | — |
| return_last_frame | BOOLEAN | true | Ask the API to return a watermark-free PNG of the last frame. Connect last_frame → first_frame of the next node to chain videos. |
| poll_interval | INT | 105–60 | Seconds between status-check requests. |
| max_wait | INT | 60060–3600 | Maximum seconds to wait for the task. |
| first_frameopt | IMAGE | Connect to enable I2V mode. Alone → first-frame mode. With last_frame → interpolation mode. | |
| last_frameopt | IMAGE | Connect with first_frame to enable first+last interpolation mode. | |
| first_frame_urlopt | STRING | HTTP URL override for first_frame. | |
| last_frame_urlopt | STRING | HTTP URL override for last_frame. | |
| reference_image_1opt | IMAGE | Enables reference mode ([Image 1] in prompt). | |
| reference_image_2opt | IMAGE | Second reference image ([Image 2] in prompt). | |
| reference_image_3opt | IMAGE | Third reference image ([Image 3] in prompt). | |
| reference_image_4opt | IMAGE | Fourth reference image ([Image 4] in prompt). | |
| ref_url_1opt | STRING | HTTP URL override for reference image 1. | |
| ref_url_2opt | STRING | HTTP URL override for reference image 2. | |
| ref_url_3opt | STRING | HTTP URL override for reference image 3. | |
| ref_url_4opt | STRING | HTTP URL override for reference image 4. | |
| reference_videoopt | VIDEO | Connect any ComfyUI VIDEO node (Load Video, another Seedance output, etc.). Takes priority over reference_video_url. Reference as [Video 1] in the prompt. | |
| reference_video_urlopt | STRING | HTTP URL of a reference video. Used only when reference_video is not connected. Reference as [Video 1] in the prompt. | |
| reference_audioopt | AUDIO | Connect any ComfyUI AUDIO node (Load Audio, etc.). Takes priority over reference_audio_url. Reference as [Audio 1] in the prompt. | |
| reference_audio_urlopt | STRING | HTTP URL of a reference audio track. Used only when reference_audio is not connected. Reference as [Audio 1] in the prompt. |
Outputs (6)
| Name | Type | Description |
|---|---|---|
| video | VIDEO | — |
| last_frame | IMAGE | — |
| frames | IMAGE | — |
| video_url | STRING | — |
| video_path | STRING | — |
| last_frame_url | STRING | — |