Recipe: Stock a WooCommerce Store from Photos

Goal: A folder of product photos and a price list become a stocked store — names, descriptions, short descriptions, and SEO written by AI from the images.

Applies to Antradus AI Publisher v2.7.0 · Last updated 2026-08-26

Ingredients

• Product photos (one good primary shot per product)
• Basic facts the camera can’t see: brand, sizes, materials, price
• A vision-capable provider: OpenAI, Claude, or Gemini (not DeepSeek)

Steps (per product — ~2 minutes each)

1. Products → Add New.
2. Set the Product image (right-hand panel) — this is what the AI will look at.
3. In the Antradus AI panel’s text box, type the invisible facts, telegram-style:
   "Brand: Acme. Sizes S-XL. 100% organic cotton. Includes carry pouch."
4. Click 📷 Look at my product photo & write everything.
5. The AI examines the photo → product name, long description, short description, SEO meta — all filled.
6. Set price/inventory, glance over the copy, Publish.

Batch rhythm that works

• Upload all photos to the Media Library first (drag-and-drop the folder) — then product creation is: pick image → paste facts → click → price → publish.
• Keep the facts file open beside you; copy-paste each product’s line.
• 30 products ≈ an afternoon, including review.

Gotchas

• The better the photo, the better the copy — clean background, one product per shot.
• Always add the brand and sizes in the text box — confident-sounding guesses about things the photo can’t show are the main thing to watch for in review.
• The Reporting page counts vision calls in your token totals, so catalog costs stay visible.