eCommerce
Bulk AI generation of product titles, descriptions, images and attributes.
The eCommerce component connects to your catalogue and rewrites it at scale: product titles and descriptions, AI-generated product images and missing attributes.
Requirements
At least one eCommerce connector must be active (Shopify, WooCommerce, Magento, PrestaShop, VTEX, OpenCart or Wix). Without it the catalogue is empty.
Step-by-step: your first bulk job
- Sync the catalogue — the connector imports products, collections, images and current field values.
- Filter what you want to work on: collection, missing description, missing images, low content score, price range, stock status.
- Select the products. Start with 20–50 to validate the output before spending credits on thousands.
- Choose the generation preset: language, tone of voice, length, which sections to include (benefits, materials, care, shipping), whether to skip products that already have images.
- Preview a couple of generated items and adjust the preset if the tone is off.
- Launch the job and follow it in the generation queue — jobs run in the background, you can leave the page.
- Review results side by side with the originals.
- Publish back to your store, in bulk or product by product.
Originals and rollback
Tuurbo keeps the original value of every field it rewrites, so you can always compare AI output with the source content and restore it.
Credit cost
| Operation | Credits |
|---|---|
| Product description | 5 per product |
| AI product image | 5 per image |
| Attribute fill | 1 per product |
Skipping products that already have images reduces the cost of a job significantly.
Product score tool
Each product gets a content score that highlights what is missing — short description, no alt text, missing attributes — so you can prioritize the products worth regenerating. Sort by score ascending and by revenue descending to find the highest-impact list.
Practical tips
- Fill the Knowledge base first: brand tone, claims you cannot make, size and material vocabulary.
- Run one collection at a time so you can measure the effect per collection.
- Re-run only what changed: filter by "missing" rather than regenerating everything.