ComfyUI Node
WAN 2.2 Generate All-in-One
Single-pass WAN 2.2 generation with prompts, frame guide insertion, sampling, and VAE decode.
WAN 2.2 Generate All-in-One
- high_model
- low_model
- clip
- vae
- start_images
- ref_image
- control_video
- images
◄positive_prompt►
◄negative_prompt►
◄width1280►
◄height704►
◄length49►
◄batch_size1►
◄fps24.00►
◄timing_modeframe►
◄resize_modecontain►
◄duplicate_policyerror►
◄pad_color0,0,0►
◄global_strength1.00►
◄start_images_strength0.85►
◄structural_repulsion_boost1.00►
◄seed0►
◄steps30►
◄cfg1.0►
◄sampler_nameeuler►
◄schedulersimple►
◄switch_step3►
◄shift8.00►
◄denoise1.00►
◄guides_json{"version":1,"guides":[]}►
CategoryWAN 2.2
Inputs (30)
| Name | Type | Default | Description |
|---|---|---|---|
| high_model | MODEL | WAN 2.2 high-noise image-to-video model, usually sampled first. | |
| low_model | MODEL | WAN 2.2 low-noise image-to-video model, usually sampled after switch_step. | |
| clip | CLIP | WAN text encoder used to encode prompts. | |
| vae | VAE | WAN 2.2 VAE used for guide encoding and final decode. | |
| positive_prompt | STRING | — | |
| negative_prompt | STRING | — | |
| width | INT | 128032–16384 | — |
| height | INT | 70432–16384 | — |
| length | INT | 491–16384 | — |
| batch_size | INT | 11–4096 | — |
| fps | FLOAT | 24.001–240 | Frames per second used when timing_mode is seconds. |
| timing_mode | COMBO | frame | Interpret guide positions as frame indexes or seconds. |
| resize_mode | COMBO | contain | How guide images are resized before VAE encoding. contain/pad preserves aspect ratio with padding. |
| duplicate_policy | COMBO | error | How to handle guide images that resolve to the same frame. |
| pad_color | STRING | 0,0,0 | RGB padding color for contain/pad resize mode. Accepts r,g,b or #rrggbb. |
| global_strength | FLOAT | 1.000–1 | Multiplier applied to every guide strength and start image strength. |
| start_images_strength | FLOAT | 0.850–1 | Strength for the optional start image sequence before global_strength is applied. |
| structural_repulsion_boost | FLOAT | 1.001–2 | High-noise structural repulsion boost between guide anchors. 1.0 disables the extra transition guidance. |
| seed | INT | 00–18446744073709550000 | — |
| steps | INT | 301–10000 | — |
| cfg | FLOAT | 1.00–100 | — |
| sampler_name | COMBO | euler | 44 options: euler, euler_cfg_pp, euler_ancestral, euler_ancestral_cfg_pp, heun, heunpp2, +38 |
| scheduler | COMBO | simple | 9 options: simple, sgm_uniform, karras, exponential, ddim_uniform, beta, +3 |
| switch_step | INT | 30–10000 | Step where sampling switches from high_model to low_model. Common WAN 2.2 workflows use 3 with 5 total steps. |
| shift | FLOAT | 8.000–100 | WAN flow sampling shift. This matches ComfyUI's normal ModelSamplingSD3/WAN shift control. |
| denoise | FLOAT | 1.000–1 | — |
| guides_json | STRING | {"version":1,"guides":[]} | Hidden serialized guide data used by the custom UI and saved in workflows. |
| start_imagesopt | IMAGE | Optional IMAGE batch inserted from frame 0. | |
| ref_imageopt | IMAGE | Optional WAN 2.2 reference image encoded as reference_latents. | |
| control_videoopt | IMAGE | Optional WAN 2.2 control video encoded into the first concat slot. |
Outputs (1)
| Name | Type | Description |
|---|---|---|
| images | IMAGE | — |