ComfyUI Node
IAMCCS Cine H3 Vision Info - Local VLM
A ComfyUI node in IAMCCS/MiniMax H3/Prompting with 27 inputs and 1 output.
IAMCCS Cine H3 Vision Info - Local VLM
- cine_linx
- clip
- image_1
- image_2
- image_3
- image_4
- cine_linx
◄analysis_modeh3_prompt_ready►
◄analysis_targetglobal►
◄context_merge_policyappend►
◄template_modeqwen3_vl_labeled►
◄image_role_1opening_frame►
◄image_target_1analysis_target►
◄image_role_2closing_frame►
◄image_target_2analysis_target►
◄image_role_3subject_identity►
◄image_target_3analysis_target►
◄image_role_4style_lighting►
◄image_target_4analysis_target►
◄analysis_max_side768►
◄seed0►
◄temperature0.20►
◄max_tokens768►
◄prompt_enhance_modecontext_only►
◄release_vlm_after_analysistrue►
◄custom_analysis_prompt►
◄user_direction►
◄prepared_analysis►
CategoryIAMCCS/MiniMax H3/Prompting
Inputs (27)
| Name | Type | Default | Description |
|---|---|---|---|
| analysis_mode | COMBO | h3_prompt_ready | 7 options: h3_prompt_ready, identity_continuity, subject_performance, scene_geography, camera_light_composition, wardrobe_props, +1 |
| analysis_target | COMBO | global | 5 options: global, local_auto, local_1, local_2, local_3 |
| context_merge_policy | COMBO | append | 2 options: append, replace |
| template_mode | COMBO | qwen3_vl_labeled | 2 options: qwen3_vl_labeled, model_default |
| image_role_1 | COMBO | opening_frame | 10 options: opening_frame, closing_frame, subject_identity, subject_performance, wardrobe_prop, environment, +4 |
| image_target_1 | COMBO | analysis_target | 7 options: analysis_target, global, local_auto, local_1, local_2, local_3, +1 |
| image_role_2 | COMBO | closing_frame | 10 options: opening_frame, closing_frame, subject_identity, subject_performance, wardrobe_prop, environment, +4 |
| image_target_2 | COMBO | analysis_target | 7 options: analysis_target, global, local_auto, local_1, local_2, local_3, +1 |
| image_role_3 | COMBO | subject_identity | 10 options: opening_frame, closing_frame, subject_identity, subject_performance, wardrobe_prop, environment, +4 |
| image_target_3 | COMBO | analysis_target | 7 options: analysis_target, global, local_auto, local_1, local_2, local_3, +1 |
| image_role_4 | COMBO | style_lighting | 10 options: opening_frame, closing_frame, subject_identity, subject_performance, wardrobe_prop, environment, +4 |
| image_target_4 | COMBO | analysis_target | 7 options: analysis_target, global, local_auto, local_1, local_2, local_3, +1 |
| analysis_max_side | INT | 768256–1536 | — |
| seed | INT | 00–4294967295 | — |
| temperature | FLOAT | 0.200–1.5 | — |
| max_tokens | INT | 768128–4096 | — |
| prompt_enhance_mode | COMBO | context_only | 2 options: context_only, enhance_visual_context |
| release_vlm_after_analysis | BOOLEAN | true | — |
| cine_linxopt | IAMCCS_SUPERNODE_LINX | — | |
| clipopt | CLIP | — | |
| image_1opt | IMAGE | — | |
| image_2opt | IMAGE | — | |
| image_3opt | IMAGE | — | |
| image_4opt | IMAGE | — | |
| custom_analysis_promptopt | STRING | — | |
| user_directionopt | STRING | Used only when prompt_enhance_mode=enhance_visual_context. | |
| prepared_analysisopt | STRING | Optional pre-generated vision description. When connected, it is used instead of running the local VLM. |
Outputs (1)
| Name | Type | Description |
|---|---|---|
| cine_linx | IAMCCS_SUPERNODE_LINX | — |