Demo limit reached
Your 2 hours of GPU time are used. Contact us to upgrade and keep generating.
No trained LoRA — generating with base model.
This LoRA has no trigger word — it may not activate properly.
No trained LoRA — generating with base model.
This LoRA has no trigger word — it may not activate properly.
No trained LoRA — generating with base model.
This LoRA has no trigger word — it may not activate properly.
How to get the best results
- Less is more. Anything the prompt describes overrides the LoRA. Name only what you actually want to change.
- Shots from behind: keep the subject as short as possible. A long facial description is what breaks her face in those angles (two words can be enough).
- Her own rooms: if the location is trained into the LoRA, tick it under "Trained locations" instead of describing the room — any description of the place replaces the background the LoRA knows.
Auto script.
Build a whole set from reference photos or from a script. Review every shot first, and generate only what you approve.
New batch
How do you want to build this set?
Who and how should it look?
More options
Style
What is she wearing, and where?
Wardrobe & props
Fields with * appear in every shot of a stage; if you leave one empty, the set's example is used. Give every piece its own colour and material — “white thin lace bra”, “white thin lace thong”.Place
Review and create
Nothing is generated yet: you review every shot next, and only the ones you approve use the GPU.
Batch
Preview
Models & LoRAs
No models yet.
No LoRAs yet.
Select your style
Tap any styles to combine them, then set each one's strength.
Choose photo area
Uploading photos…
Model
Training from photos is not available yet. For now, import an existing .safetensors below to link a LoRA to this model.
Drop all your photos at once, or use + Add images — captions and tags are written for you a few seconds after upload, so there is nothing to fill in. Flagged photos tell you in one line what to fix. Aim for 15–60 photos total.
Always written into every photo of this identity — piercings, tattoos, a beauty mark. Small things that need spelling out.
Import an existing .safetensors to link it to this model (skips training).
Training runs on our servers — you can keep working or close this, it won't stop. We'll have it ready here when it's done.
LoRA
Always written into every photo of this identity — piercings, tattoos, a beauty mark. Small things that need spelling out.
Your tasks
No tasks assigned to you.
No tasks yet.
Profile
Beta Preview
Connections
Loading…
Loading…
Loading…
Loading…
Claude
Loading…
Instagram & TikTok Add to Swap
Capture details
Sign out
Members of your workspace
No members yet.
No activity yet.
No videos yet.