“Upload a photo, get a listing” sounds like magic, and when it works it feels like it. But it isn't magic — it's a sequence of specific steps, and knowing them tells you both why the output is usually good and exactly when to step in. Here's what actually happens when Kiln Vision turns a product photo into an Etsy listing, and the handful of cases where a single line of context from you makes the difference.
Upload one clear product photo and you get back the three parts of a listing that decide whether buyers find it: a title written to Etsy's April 2026 guidance (item first, top traits up front, under 15 words), all 13 tags, and a full description — plus a 0–100 score on how strong the result is. You review and edit; nothing publishes on its own. The value isn't just speed, it's that the photo already contains the colour, material and style Etsy wants front-loaded, so the hardest part of writing — noticing and naming those details — is done before you type a word.
First the model reads the image the way a careful person would, building an inventory: what the product is, how many items are in the shot, its colour, silhouette and material, and its likely use. It's told to describe only what it can actually see and to stay literal — a photo of a matte black ceramic mug becomes “matte black ceramic mug,” not an invented backstory. That literal read matters, because “matte black ceramic mug” is exactly the specific phrase a ready-to-buy shopper types, while “black mug” reaches almost no one.
A list of visual details isn't yet a listing that ranks. So the next step grounds those details in demand: live buyer search suggestions from Google (US), ListifyAI's own rolling keyword-demand history, and a seasonal calendar that knows which occasions are near. The honest label matters here — the suggestions come from Google, not from a private view of Etsy's search bar. That grounding is what turns “ceramic mug” into the phrases buyers are actually searching this month, and it's the part a plain image description can't do on its own.
Etsy gives you 13 tags, each up to 20 characters, and reads every one — so the generator fills all 13, never fewer. A ladder builds them: it takes the strongest candidate phrases, keeps only ones that fit inside 20 characters and use at least two words, and de-duplicates so no two tags are the same search (it even catches near-duplicates and simple reorderings). If the pool comes up short, a second pass tops it up. The result is 13 distinct buyer phrases that expand your reach instead of repeating your title.
After the listing is written, a second pass compares the finished text back against the photo and reports what's still missing — a match score plus the specific buyer signals your words don't yet mention. If the photo clearly shows a gift-ready box or a colour variation the description skipped, this is where it surfaces. It's the step that separates a generator from a chatbot: not just writing from the image, but checking the writing against the image and telling you where to add.
Three situations are worth watching. Branded items: the model deliberately won't guess a specific model name it can't verify, so if the brand or model is the selling point, type it in. Multi-product or bundle photos: it counts what it sees, but a busy flat-lay can confuse what's for sale versus what's a prop, so a line like “this listing is the mug only” helps. And unusual materials or techniques a photo can't reveal — a specific clay, a hidden mechanism, a personalisation option — should be added by you. In every case the fix is one sentence of context, not a rewrite.
Because everything starts from the image, the photo does most of the work of quality. Shoot on a clean, uncluttered background so the product is unmistakable. Use good natural light — accurate colour is how the model names the material correctly. Get the whole product in frame, and add a close-up if the texture or craftsmanship is the selling point. One product, clearly lit, plain background: that's the difference between a specific listing and a safe, generic one. For more on what the AI sees versus what buyers see, read what your Etsy photo tells buyers and turning visual signals into tags.
Photo-to-listing is on the free plan: 3 full listings a month, no card, photo or text. Paid plans lift the daily limits and add extras — one of them being product-preserving scene photos, which stage your saved product image in six room settings (5 a month on Starter, 20 on Pro, 60 on Agency; the free plan doesn't include scene photos). The listing writing itself, though — title, 13 tags, description and the photo-to-text check — is what you get on every plan, free included.