ComfyUI Node
LTX Reference Tags
Assembles '<Image 1> is the woman. <Image 2> is the man. <scene>'. Order matters: tag N must describe the image wired into reference_N on the slots node, and must match the convention the checkpoint was trained with. Describe the subjects DISTINCTLY — two references both called 'the woman' carry no information about which is which.
LTX Reference Tags
- prompt
◄subject_1►
◄subject_2►
◄subject_3►
◄subject_4►
◄subject_5►
◄subject_6►
◄scene►
CategoryBFS/LTX
Inputs (7)
| Name | Type | Default | Description |
|---|---|---|---|
| subject_1opt | STRING | What reference image 1 is, as a short noun phrase — 'the woman', 'the red sports car'. Becomes '<Image 1> is <this>.'. Leave empty to skip. | |
| subject_2opt | STRING | What reference image 2 is, as a short noun phrase — 'the woman', 'the red sports car'. Becomes '<Image 2> is <this>.'. Leave empty to skip. | |
| subject_3opt | STRING | What reference image 3 is, as a short noun phrase — 'the woman', 'the red sports car'. Becomes '<Image 3> is <this>.'. Leave empty to skip. | |
| subject_4opt | STRING | What reference image 4 is, as a short noun phrase — 'the woman', 'the red sports car'. Becomes '<Image 4> is <this>.'. Leave empty to skip. | |
| subject_5opt | STRING | What reference image 5 is, as a short noun phrase — 'the woman', 'the red sports car'. Becomes '<Image 5> is <this>.'. Leave empty to skip. | |
| subject_6opt | STRING | What reference image 6 is, as a short noun phrase — 'the woman', 'the red sports car'. Becomes '<Image 6> is <this>.'. Leave empty to skip. | |
| sceneopt | STRING | The scene description, appended after the declarations. Refer back to the subjects by tag here ("<Image 1> hands <Image 2> a cup") — a tag that only ever appears in the declaration has to carry both jobs at once, which is a harder association for the model to use. |
Outputs (1)
| Name | Type | Description |
|---|---|---|
| prompt | STRING | — |