Fillman converts photos to products.
Fast and cheap.
Take photos of your products and upload them exactly as they are. Unnamed, unsorted. Fillman groups them into the right products, writes every product detail, and publishes to Shopify. Point it at one product or your whole catalog - it's a bulk import tool, not just a single-listing trick.
See exactly what Fillman generates
Real Shopify admin, real fields. Here's what comes from your photos.
Illustrative mockup of a real Shopify admin page, not a live screenshot. The catalog count and the numbers below are real.
7,422 photos → 1,586 live products. Under 6 hours.
Every product on CoasterHouse started as a raw phone photo - no studio shoot, no pre-sorted files. Identifying the design, writing SEO copy, matching the right image to the right frame-color variant, and publishing it to Shopify used to mean a spreadsheet, a bulk-import tool, and a person doing the recognition and data entry by hand.
Our own pipeline did the full 7,422-photo, 1,586-product run in 4-6 hours of actual processing time.
The manual alternative: at a conservative 10 minutes per product - sorting the photo, identifying what it actually is, writing the title and description, mapping the right image to the right variant - that's 264 hours of labor. Outsourced at $5-8/hour, that's $1,300-2,100 in cost alone, plus 6-10 weeks of calendar time, before counting the rework every low-cost freelancer batch needs when a design gets misidentified or a variant image gets mismatched.
Based on CoasterHouse's actual raw photo archive (7,422 source images) and live product count (1,586) as of September 2026. Automated processing time is our own measured pipeline run. Manual-labor estimate assumes a 10 min/product average across identification, copywriting, and variant-image matching; actual outsourced quality/reliability at the low end of this rate range is inconsistent in practice.
Design one product. Fillman builds the rest like it.
You set up a single reference product exactly the way you want it - fields, metafields, variant structure. Every new photo that matches its type gets built the same way, automatically.
1. You design one
Build a single product exactly the way you want it to look - title style, fields, metafields, variant layout.
2. Fillman reads the shape
That product becomes the canonical template for its product type - not a rigid form, a real example.
3. The rest follows it
New photos that match its type get the same structure applied automatically, product after product.
Beyond title and description
The rest of the page, and the rest of the catalog, work the same way - faster than doing it by hand.
Metafields, filled in
Materials, care instructions, size charts, whatever custom fields your template uses - Fillman populates those too, not just title and description.
Your own templates
Fillman fills the fields your existing Shopify theme and templates already expect. No rebuild, no new theme.
Apply it in bulk
Run the same population across your whole catalog at once, or scope it to a specific product type - your choice, every time.
Variant images, mapped
Color, size, and style variants get matched to the right image automatically - no manual re-assigning after import.
We used Matrixify. Here's what we found.
Real testing, on a real Shopify store, before we built Fillman. Not a hypothetical.
We don't have a precise speed multiplier to quote yet - only the two differences above, which we saw directly.
How it works
Three steps, no spreadsheet in sight.
1. Add your photos
Drop in a folder of product photos, however messy. No sorting or renaming first.
2. Fillman sorts and writes
Matches each photo to the right product and variant, orders the images, and writes the listing.
3. Review and ship
Low-confidence matches get flagged for your review instead of guessed - you push the rest straight to Shopify.
We tested this properly
No cherry-picked demo. Our own lab results, scored one by one, including the ones we got wrong. Don't take our word for it - come test it yourself.
12 sellers, 12 phrasings, one right answer
A coaster with a horror-movie clown got titles from “Bill Skarsgård fan art” to “Sketch Card”. Fillman recognized all 12 as Pennywise. Confidence: 1.00.
IKEA's own sofa, and Fillman said so
Fed a photo of the EKTORP sofa, 9 of 12 matches said “UPPLAND” instead. Fillman scored it 0.08 and flagged for review instead of guessing.
Tested outside our own catalog too
Auto-apply rate at the confidence threshold that still gives 100% accuracy, per sector.
Where we're honest about our limits
Fillman struggles with generic minimalist jewelry and generic workwear apparel. When it's unsure, it flags the item for review instead of guessing - by design.
Not just a photo tool
Photo-to-listing is our starting point. The goal is a full Shopify data tool: fast bulk import/export, real field mapping from your own spreadsheets, and a UI that doesn't require a developer to operate.
Join the waitlist
We're onboarding a small number of beta sellers first. Tell us what you sell and we'll reach out.