All models
ControlNet Pose
Generate an image from a prompt while preserving a detected human pose.
Edit & Analyze Imagev1.0.010 inputsIMAGEIMAGE


Prompt: “a fantasy knight in ornate silver armor”
Every port this node publishes, with its type, its hint and the default it runs with. controlnet-pose is the id you call.
| image* | IMAGE | Reference image | |
| prompt* | TEXT | ||
| imageResolution | TEXT | 256 · 512 · 768 | 512 |
| detectResolution | INT | Pose detection resolution | 512 |
| steps | INT | 20 | |
| guidance | FLOAT | 9 | |
| additionalPrompt | TEXT | best quality, extremely detailed | |
| negativePrompt | TEXT | longbody, lowres, bad anatomy, bad hands, missing fingers, extra digit, fewer digits, cropped, worst quality, low quality | |
| eta | FLOAT | Noise | 0 |
| seed | INT |
| image | IMAGE | Generated image | |
| pose | IMAGE | Detected pose |