ComfyUI Node
Bernini Prompt Enhancer
Build the official task-aware Bernini prompt-enhancement instructions. Connect system_prompt and user_prompt to an external LLM node, then parse its response.
Bernini Prompt Enhancer
- source_video
- source_image
- reference_images
- system_prompt
- user_prompt
- llm_prompt
- api_prompt
- json_mode
◄prompt►
◄task_typev2v►
◄video_frames3►
Categoryconditioning/video_models
Inputs (6)
| Name | Type | Default | Description |
|---|---|---|---|
| prompt | STRING | Original user instruction or generation prompt. For Chinese input, the enhancer asks the external LLM to rewrite it in English. | |
| task_type | COMBO | v2v | Bernini task type. Choose the mode that matches connected media and the intended generation or editing workflow. |
| video_frames | INT | 31–8 | Number of source video frames described to the external vision LLM in prompt-enhancement tasks. |
| source_videoopt | IMAGE | Optional source video frames. Also connect the same frames to your external vision API node when the selected task needs visual context. | |
| source_imageopt | IMAGE | Optional source image. Also connect it to your external vision API node for i2i/i2v tasks. | |
| reference_imagesopt | IMAGE | Optional reference image batch. Also connect it to your external vision API node for r2v/r2i/rv2v/vrc2v/vi2v tasks. |
Outputs (5)
| Name | Type | Description |
|---|---|---|
| system_prompt | STRING | — |
| user_prompt | STRING | — |
| llm_prompt | STRING | — |
| api_prompt | STRING | — |
| json_mode | STRING | — |