ComfyUI Node
BFS Shot H3 Conditioning
Build the usual MiniMax H3 shot conditioning with references, prompt and an aligned guide and/or native reference video. Use the Advanced node for duet, shifted RoPE and mask-guide tools.
BFS Shot H3 Conditioning
- shot
- clip
- vae
- audio_vae
- vlm
- positive
- latent
- prompt
◄guide_modealigned guide (Add Guide)►
◄use_ref_2true►
◄ref_image_sizematch►
◄with_audiofalse►
◄inpaintper shot (planner)►
◄source_bluroff►
◄source_blur_strength1.00►
CategoryBFS/shot loop
Inputs (12)
| Name | Type | Default | Description |
|---|---|---|---|
| shot | BFS_SHOT | — | |
| clip | CLIP | — | |
| vae | VAE | — | |
| guide_mode | COMBO | aligned guide (Add Guide) | aligned guide: preserve the shot frame by frame with MiniMax H3 Add Guide. native reference video: provide it as <Video 1> context. both: use both paths. |
| use_ref_2 | BOOLEAN | true | Also pass the second reference. |
| ref_image_size | COMBO | match | 2 options: match, max |
| audio_vaeopt | VAE | — | |
| with_audioopt | BOOLEAN | false | Condition on the shot soundtrack too. This can improve lip sync when audio_vae is connected. |
| vlmopt | CLIP | Optional Qwen3-VL encoder. Fills missing reference details in the planner prompt. | |
| inpaintopt | COMBO | per shot (planner) | Use each shot's Planner mask mode, force mask-only generation, or disable it. |
| source_bluropt | COMBO | off | Blur the source subject in the guide so the model is less likely to copy its identity. |
| source_blur_strengthopt | FLOAT | 1.000.1–1 | — |
Outputs (3)
| Name | Type | Description |
|---|---|---|
| positive | CONDITIONING | — |
| latent | LATENT | — |
| prompt | STRING | — |