LoraTag

LoraTag is an image captioning tool built for preparing datasets to train LoRA models. It helps you label hundreds of images in minutes and produce structured, detailed captions.

Faster LoRA dataset prep

Upload individual images or entire directories, and the model generates captions for each file at the detail level you choose. Batch processing keeps large dataset labeling in a single run, without repetitive manual steps.

Control caption detail and tags

You can tune how detailed the descriptions are to match your training goal—from broad tags to more specific scene, style, and object descriptions. This makes it easier to iterate on LoRA training and improve the quality of generated results.

Who it’s for

• Artists building custom LoRA datasets
• Developers and hobbyists training LoRA models
• Anyone labeling large image collections who wants consistent captions