이커머스용 AI 상품 이미지 API: gpt-image-2, 장당 ¥0.16

Capabilities

Pricing and cost comparison

The API is billed per image with no minimum spend. Here it is next to traditional methods so engineering / procurement can do the math:

Costs are estimates for order-of-magnitude only; the console billing is authoritative. Traditional prices vary widely by category / region.

Quickstart

With an API key, one curl generates your first image (replace YOUR_API_KEY with the key from your console):

curl https://api.tutujin.com/v1/images/generations \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-image-2",
    "prompt": "A skincare bottle on a clean pure-white background, e-commerce main image",
    "size": "1024x1024",
    "n": 1
  }'

The response includes an image URL; for image-to-image, pass the reference image too. Full parameters (size / n / image-to-image) are in the console API docs.

Use cases

자주 묻는 질문

Does Tutujin have an AI product image API?

Yes. Tutujin offers a gpt-image-2 REST image-generation API at about ¥0.16 per image, with batch concurrency and image-to-image, built for e-commerce teams to wire into ERP / listing systems for bulk product images. Sign up and open a key in the console to call it.

How is the API priced — what does one image cost?

gpt-image-2 is about ¥0.16 per image, billed by images actually generated, with no minimum spend. Versus ¥50–200/SKU studios or ¥10–80/image retouching, bulk generation can cut cost by 90%+.

How do I get an API key?

Sign up for a Tutujin account, then open an enterprise / developer API key in the console and send it in an HTTP header. The key is only visible after login and never shown on public pages.

Can I batch-generate and integrate with my system?

Yes. The endpoint supports multiple images per request (n) plus concurrent submission, so you can queue thousands of SKUs, and you can embed it directly in an ERP / product backend / listing tool to generate main images per SKU.

Which model is it, and does it support image-to-image?

The model is gpt-image-2, supporting both text-to-image and image-to-image. For image-to-image, upload one product reference and it swaps background / white bg / scene while keeping the subject — great for many versions of one product.

Who is it for, and how does it differ from the web app?

The API suits e-commerce teams with developers who want to embed generation in their own system or pipeline. Individuals / small sellers are better off with the web app (sign up, pick a scene, pay with credits) — no code needed.