Control product images, not another prompt lottery.
ProductJSON Studio turns product photos, references and listing briefs into structured JSON prompts for ecommerce images. It is built for sellers who need consistent logos, shapes, packaging, lighting and marketplace formats.
Agent-assisted setup with Claude Code or Codex. Bring your own provider keys.
From scattered prompts to one structured JSON.
Product-image hallucination is not a creative issue. It is a control issue.
Ecommerce images fail when the model changes the logo, shape, material, package size, included accessories or platform format. ProductJSON Studio makes those parts explicit fields instead of hidden assumptions inside a plain-text prompt.
Reference library
3,004 reviewed ecommerce references, plus 240 included generated 2K marketplace-reference examples.
JSON control
Change product, lighting, background, camera, or platform fields without rebuilding the whole image idea from scratch.
Product identity
Use product reference locking to reduce drift in logo, shape, colour, material, packaging and proportions.
Model-eval routing
Compare Nano Banana 2 / Pro against GPT Image 2 and choose the better model for the reference style.
240 generated 2K examples, built around the images sellers actually need.
The upgrade is not just more images. It is a structured set of ecommerce reference formats across 30 product types, generated through Magnific with Nano Banana Pro and QA'd for product clarity, marketplace usefulness and prompt-learning value.
Non-technical setup path
Buyers open the source folder in Claude Code or Codex, paste the included setup prompt, then let the agent install dependencies, run checks and start the local app. Provider keys stay with the buyer and are pasted into Settings only when they are ready to create.
- Download the source-code ZIP from Gumroad.
- Open the folder in Claude Code, Codex, Cursor or another coding agent.
- Paste the included setup prompt and let the agent install, test and start the local app.
- Open the local URL, browse Learn/Inspiration first, then add your own provider keys in Settings when you are ready to create.
- Start with the Learn section, then create your first Amazon or Shopify image set.
What is included
- Browse the reference library to find the marketplace image format you need: main image, clean product shot, lifestyle, exploded view, callout, scale, box contents or comparison.
- Convert an image, URL or plain prompt into structured JSON so product, scene, lighting, camera, platform rules and negative constraints are editable fields.
- Lock the product reference so the logo, shape, colour, material and proportions stay anchored while you change the background or image type.
- Use the model-eval workflow to decide whether Nano Banana 2 / Pro or GPT Image 2 is the better fit for the reference style.
- Generate Amazon and Shopify image sets from one product brief, starting with the main image before building the secondary gallery.