This LoRA turns just about anything into a toys and action figures, complete with articulated joints, realistic plastic materials and painted details.
The LoRA is based on Qwen Image Edit 2511 and is trained on a variety of photos and illustrations, to make it as flexible as possible.
The LoRA is super simple to use - simply load a target image into the provided
workflow and use the trigger phrase "action the [thing you want to turn into a toy/action figure]". For automation purposes you can also use the phrase "action the scene", which works in most cases.
Experimenting with different sampler/scheduler combos can also give improved and/or unique results, so we encourage you to try some variations out for yourself. Some examples of combos we've seen good results with: euler/simple, ddim/beta57 and er_sde/bong_tangent
Prompt: Action the character
Prompt: Action the tennis player
