by SizeGuide (LX AI)
SizeGuide turns product measurements and fit notes into clearer size-chart copy — so shoppers guess less and merchants see fewer wrong-size returns.
Decision-support only — confirm before you publish or act.
I started SizeGuide after seeing too many returns caused by vague size charts. The tool takes raw data that a merchant already has — body measurements, garment specs, fit notes from designers — and turns that into plain‑language copy that appears on the product page. The flow is simple: upload a spreadsheet, pick a template, review the generated text, then publish. Each step is designed to keep the merchant in control while reducing the guesswork for shoppers.
First, the merchant supplies a CSV with columns for chest, waist, hip, inseam, rise, and any fit comments such as “runs large” or “tight at the thigh”. The system validates the data, flags missing values, and suggests defaults based on the product category. Next, the merchant chooses a copy style — bullet list, paragraph, or hybrid — and selects a tone (neutral, friendly, technical). The engine then applies a set of rules that map each measurement to a size label (XS, S, M, L, XL, XXL) and writes a sentence that explains what the number means for fit.
Finally, the merchant sees a preview of the size‑chart block exactly as it will appear on the storefront. They can edit any sentence, reorder bullets, or add a custom note. When satisfied, they click “Publish” and the HTML snippet is injected into the product template. Because the tool only suggests copy, the merchant must confirm before the change goes live.
Each field accepts numeric values. The system expects one row per size variant. If a merchant provides only a subset of fields, the engine fills the gaps using industry averages for that garment type, but it highlights those estimates in the preview so the merchant can override them.
Free‑text comments are captured in a separate column. Typical entries include:
The parser looks for keywords such as “small”, “large”, “tight”, “loose”, “true”, and converts them into modifiers that adjust the wording of the generated size‑chart sentences.
The core of SizeGuide is a deterministic rule engine. It does not use machine learning to guess intent; instead it follows a hierarchy of if‑else statements that I wrote based on years of fitting sessions. For example:
These rules are stored in a JSON file that merchants can view but not edit directly. The reason is to keep the output predictable and avoid unexpected phrasing that could confuse shoppers.
Three template styles are available:
The merchant selects a template in the UI. The engine then fills placeholders like {{size}}, {{range}}, {{note}} with the values derived from the rule set. Because the templates are static HTML snippets, the resulting copy loads fast and does not rely on external scripts.
After the engine runs, a modal shows the generated markup. The merchant can:
All edits are saved locally until the merchant confirms. The preview uses the exact CSS classes that the storefront theme expects, so the look matches the live site without extra styling work.
| Size | Generated copy (bullet style) |
|---|---|
| XS | Fits chest 30‑32 in, waist 24‑26 in, hip 34‑36 in – true to size. |
| S | Fits chest 33‑35 in, waist 27‑29 in, hip 37‑39 in – runs slightly large in the waist. |
| M | Fits chest 36‑38 in, waist 30‑32 in, hip 40‑42 in – true to size. |
| L | Fits chest 39‑41 in, waist 33‑35 in, hip 43‑45 in – runs small, consider sizing up. |
| XL | Fits chest 42‑44 in, waist 36‑38 in, hip 46‑48 in – true to size. |
| XXL | Fits chest 45‑47 in, waist 39‑41 in, hip 49‑51 in – generous cut, looser fit. |
The table above shows a typical result for a men’s casual shirt. Notice how each line includes the measurement range and a fit note derived from the source data.
SizeGuide outputs a plain HTML snippet that works with any platform that allows custom liquid, HTML blocks, or schema JSON. I have tested it on:
No app installation is required; the merchant copies the generated code and pastes it where the size chart normally appears.
The snippet includes semantic markup that helps search engines understand the content. Each size line uses the dl element with dt for the size label and dd for the description. This pattern aligns with the schema.org QuantitativeValue pattern and can be picked up by Google’s product rich results.
Because the tool only provides copy, the merchant must still ensure that the overall product page meets Google’s guidelines for structured data. I recommend validating the final page with the Rich Results Test before publishing.
The generated HTML is under 5 KB for a typical size chart with six sizes. No external scripts, no fonts, no tracking pixels are added. Page load time increases by less than 5 ms on a mid‑range connection, which is negligible compared to image loading.
I want to be clear about the limits of SizeGuide so you can decide if it fits your workflow.
First, SizeGuide is not a guarantee that returns will drop to zero. It reduces guesswork, but factors such as photography, model pose, and customer expectations still influence fit perception.
Second, the tool does not replace a professional fit session or a wear‑test panel. It works best when you already have reliable measurements and honest fit notes from designers or sample makers.
Third, SizeGuide does not generate images, videos, or interactive fit finders. It only produces textual copy that sits alongside your existing visuals.
Fourth, the engine is not a black‑box AI model that learns from your data. It is a rule‑based system that you can inspect, audit, and modify only through the provided JSON rule file. This makes the output predictable and easier to troubleshoot.
Fifth, SizeGuide does not handle legal compliance for accessibility or language localization. You must review the generated copy for WCAG contrast if you apply custom styles, and you must translate the text yourself if you sell in multiple languages.
Finally, the tool does not store your measurement data beyond the current session. Once you leave the builder, the CSV is discarded unless you download it. I built this privacy‑first approach because I know how sensitive product specs can be.
If you are looking for a solution that promises 100 % accuracy, eliminates all returns, or automatically creates visual fit guides, SizeGuide is not that tool. It is a decision‑support aid that makes the copy‑writing step faster and more transparent.
I keep a written checklist next to every release that touches this topic. The list is boring on purpose: name the owner, name the evidence artifact, and name the rollback. If a step cannot produce a file, a ticket, or a log line, it is not a step — it is a wish.
| Signal | What to record | Stop-ship? |
|---|---|---|
| Unexpected write or send | Tool name, args, identity | Yes until gated |
| Missing confirmation on destructive action | Call path and bypass | Yes |
| Untrusted text treated as instruction | Source surface and resulting action | Yes for agent reach |
| Scanner clean, judgment open | Deferred items + owner | Only if core path safe |
sizeguide helps with the measurable first pass. Architecture choices, legal sign-off, and customer-facing claims still need a human owner. Treat every checklist row as decision support, not a certificate.
When a prospect asks whether the product "covers everything," I answer with the boundary: automated checks catch known patterns; novel chains and policy judgment still need review. That honesty keeps support load down and keeps the blog aligned with what the product actually does.
If you need a second opinion, pair this write-up with the references below and with your own staging reproduction. Do not skip the reproduction step just because a model wrote a confident summary.
SizeGuide turns product measurements and fit notes into clearer size-chart copy — so shoppers guess less and merchants see fewer wrong-size returns.
Decision-support only — confirm before you publish or act.
Want to try the product? SizeGuide
Generated 2026-09-17 · primary sources only · no fabricated traffic metrics