RunComfy
ComfyUI
Trainer
New
Models
API
Pricing
EN
Sign In
TRAINING & INFERENCE
Train LoRA
Datasets
LoRA Assets
Run LoRA
Generations
DEDICATED ENDPOINTS
Deployments
Requests
ACCOUNT
Usage
API Docs
API Keys
How-tos
Run LoRA
runcomfy/krea2-o-edit
runcomfy/krea2-o-edit
Playground
API
LoRAs
*
LoRA #1
Path
*
Loading LoRAs...
This list comes from
LoRA Assets
. Upload LoRAs there to see more options.
LoRA Scale
Add Item
Prompt
*
Control Image 1
*
Control Image 2
Control Image 3
Width
Height
Guidance Scale
Sample Steps
Negative Prompt
Seed
Sampler
KV Cache
Must match the training config's model_kwargs.kv_cache. It changes the reference attention mask, so a mismatch silently produces wrong output. Turn off for adapters trained before 2026-07-16.
Match Target Resolution
Must match the training config's model_kwargs.match_target_res. On, each reference image is area-matched to the target resolution; off, it is only capped at 1 MP.
Use Training Settings
Generate
Idle
Your request will cost
$0.045
per megapixel of generated image.
Click Generate to start
Results will appear here
