LoRA Fine-Tuning API — Train Custom Models

Train a custom LoRA fine-tune on your own images via the Stable Diffusion API. From $3 per training job.

LoRA Fine-Tuning

Train a custom LoRA on your own image dataset. Output is a portable LoRA file you can use with any compatible Stable Diffusion model.

Pricing

LoRA training starts at $3 per training job.

Request

curl -X POST 'https://stablediffusionapi.com/api/v3/lora_finetune' \
  -H 'Content-Type: application/json' \
  -d '{
    "key": "YOUR_API_KEY",
    "instance_prompt": "photo of sks dog",
    "class_prompt": "photo of a dog",
    "base_model_type": "sd-1.5",
    "negative_prompt": "low quality",
    "training_images": ["https://example.com/img1.jpg", "https://example.com/img2.jpg"],
    "max_train_steps": "2000",
    "lora_type": "lora"
  }'

Use the returned training_id to poll for completion.