ComfyUI Node
Wan Adaptive Flow
Creates adaptive multi-pass video generation with improved transitions
Wan Adaptive Flow
- text_embeds
- model
- clip_vision
- vae
- image_embeds
- t5_encoder
- text_embeds
- bridge_info
◄analysis_points3►
◄first_pass_steps8►
◄bridge_modeAutomatic►
◄bridge_count1►
◄bridge_strength0.75►
◄lighting_weight0.80►
◄content_weight0.70►
◄composition_weight0.60►
◄save_first_passtrue►
◄original_prompts—►
CategoryWanSeamlessFlow
Inputs (16)
| Name | Type | Default | Description |
|---|---|---|---|
| text_embeds | WANVIDEOTEXTEMBEDS | — | |
| analysis_points | INT | 31–10 | Number of points to analyze per transition |
| first_pass_steps | INT | 81–30 | Number of steps for initial generation pass |
| bridge_mode | COMBO | Automatic | Type of bridge prompt to create |
| bridge_count | INT | 11–5 | Number of bridge prompts per transition |
| bridge_strength | FLOAT | 0.750.1–1 | How strongly to apply bridge modifications |
| modelopt | WANVIDEOMODEL | WanVideo model for initial pass | |
| clip_visionopt | CLIP_VISION | CLIP Vision for frame analysis | |
| vaeopt | WANVAE | VAE for decoding latents | |
| image_embedsopt | WANVIDIMAGE_EMBEDS | Image embeddings for initial pass | |
| t5_encoderopt | WANTEXTENCODER | T5 encoder for bridge prompts | |
| lighting_weightopt | FLOAT | 0.800–1 | — |
| content_weightopt | FLOAT | 0.700–1 | — |
| composition_weightopt | FLOAT | 0.600–1 | — |
| save_first_passopt | BOOLEAN | true | — |
| original_promptsopt | STRING | Original prompts separated by | |
Outputs (2)
| Name | Type | Description |
|---|---|---|
| text_embeds | WANVIDEOTEXTEMBEDS | — |
| bridge_info | STRING | — |