Give a generic AI listing tool a photo of a Patagonia Better Sweater and it can come back with 100% recycled merino wool, made in Italy, in a colorway the brand never produced. None of that is on the tag, and no buyer searching for the actual piece will ever match it.
Generic AI tools have a structural problem. They generate text from a category prompt and a few photos. They don't have the actual fabric percentages on the inside-collar tag, the real measurements off the cloth, or the seller's shipping policies. The output reads as plausible to someone who doesn't know the garment, which is the audience least likely to convert.
AI Descriptions works from real garment data. The output draws on the photo you captured, the label you scanned, the measurements you took, and the seller profile you set up once. Every claim traces to something the app already verified. The descriptions read as informed because they are.
What follows is what the generator actually does, the 9 marketplaces it tunes output for, and how to get descriptions that convert.
What AI Descriptions does
The generator takes the garment record (photo, label data, measurements) and writes a listing description tuned to the platform you're listing on. Brand name from the scanned label. Fabric percentages from the same scan. Measurements from the capture. Care instructions from the same tag.
Descriptions are truthful, accurate, and conversion-optimized in that order. Truthful because the source data is real garment data. Accurate because the percentages, measurements, and care instructions match the actual piece. Conversion-optimized because the output is tuned per platform: keyword density for Etsy, hashtag count for Depop, character ceilings for Poshmark and Mercari.
Cross-platform mode generates a single description that respects every supported platform's char limit and tone. Single-platform mode targets one specifically.
The 9 marketplaces
Each platform has its own char limit, tone, and hashtag convention. The generator respects all three.
| Platform | Char limit | Tone | Hashtags |
|---|---|---|---|
| eBay | Unlimited | Professional | No |
| Poshmark | 1,500 | Casual / friendly | Limited |
| Mercari | 1,000 | Casual | Light |
| Depop | Free-form | Casual / Gen-Z | Heavy |
| Grailed | Free-form | Luxury / streetwear | No |
| Etsy | Free-form | SEO keyword-heavy | Yes |
| Shopify | Free-form | E-commerce / professional | No |
| Amazon | Free-form | Bullet-led, keyword-dense | No |
| Online Store mode | Free-form | Neutral, generic e-commerce | Selective |
Online Store mode is the cross-platform fallback for fashion brands running their own storefront on a non-supported platform. The output reads as neutral e-commerce copy that works on Webflow, Squarespace, BigCartel, and most direct-to-consumer site builders.
How to generate a description
The flow assumes a completed capture and a label scan. Both are recommended; descriptions improve materially when label data is included.
- 1
Capture the garment
Follow the flow in our capture guide. Measurements feed the description directly.
- 2
Scan the label
Brand, fabric composition, stretch, care instructions all flow into the description body. See the Label Scanner guide.
- 3
Tap "Generate Description"
From the Results screen, select the marketplace, and the generator writes a tailored description in 5 to 10 seconds.
- 4
Set up the seller profile
A one-time configuration in Settings: shop name, shipping policy, return window, custom listing notes. Every future description respects these defaults.
- 5
Review and export
The description shows in a preview pane with a character count. Edit anything, then copy it or share it straight into your listing.
End to end, generating a description for a single platform takes 15 to 30 seconds after capture and label scan.
Why these descriptions actually convert
Listings convert when they answer the buyer's question without a follow-up message. Buyers ask about fit, fabric, condition, care, shipping, and returns, in roughly that order. A Size AI description answers the first four out of the box because the source data is the actual garment. Generic AI tools answer none of those reliably because they don't have access to the data. The conversion delta shows up in days-to-sale.
Tips that actually move the needle
- Scan the label first. Descriptions without label data fall back on the photo and measurements only, which means fabric content and care instructions get omitted. Both fields convert disproportionately well.
- Set up the seller profile once. Shipping policy, return window, and shop voice all persist across listings.
- Add a condition note before generating. A two-sentence disclosure (small fade on the right cuff, taping on the inside seam still intact) trains the generator to include it in the right spot.
- Use cross-platform mode for inventory listed in three or more places. A single generated description that respects every platform's limits saves the per-platform regenerate.
Limitations
- The 20 description languages are an iOS app feature. The website is English only.
- Poshmark (1,500 chars) and Mercari (1,000 chars) have hard character truncation. The generator respects those limits automatically.
- The generator does not invent fabric percentages, brand details, or care instructions. If the source data is missing a field, the generator omits the corresponding sentence rather than filling in a guess.
