Make every size
the right size
SizeGuide is the sizing layer for modern commerce. Verified charts, explainable recommendations, and APIs built for stores, shoppers, and AI agents.
Free tier for merchants and developers · No order-volume surcharges
Classic Cotton Tee
We recommend M for you
Explainable by design · see how this works
of recommendations ship with reasoning, confidence, and chart provenance
Methodologyprovenance tiers on every chart: official, published, or reference. No fourth state
Provenance standardShoppers don't hesitate over price.
They hesitate because
they aren't sure it fits.
Fit uncertainty is the last unanswered question on the product page. SizeGuide exists to answer it, with evidence.
One sizing engine.
Every surface it touches.
Charts, recommendations, and analytics run on the same deterministic core, so your product page, your data, and your AI channels always agree.
Recommend with confidence
Body-based size recommendations with per-zone fit, honest confidence scores, and an alternative when the shopper sits between sizes.
Explore recommendationsCharts with provenance
Versioned, dated, sourced size charts. Import from CSV, metafields, or a photo, with review before anything reaches your storefront.
Explore smart charts{ "recommended_size": "M",
"confidence": 0.92,
"explanation": { … }
}
Integrate anywhere
REST endpoints, an MCP server, and a one-line embed. The same engine your PDP uses, exposed to your stack and to AI agents.
Read the docsBuilt for everyone sizing touches
Install in minutes
Connect your store, import charts from CSV, metafields, or a photo, map to products, publish. No developer required.
Prove the lift
Honest A/B against control from day one: conversion after recommendation, size distribution, return-risk flags.
Pricing that scales with you
Free, $29, $99, and enterprise. Flat tiers with no order-volume surcharges.
Answers, not size charts
A recommendation with reasoning: your measurements, the chart row, the fit per zone.
Between sizes? We say so
Lower confidence is shown, not hidden. You see the alternative size and when to pick it.
Private by default
Measurements stay yours. Merchants see a size, never your body data. Deletion is one click.
One contract
Typed engine interface, versioned charts, golden-vector test corpus. Same inputs, same output, every surface.
REST + MCP
Stateless endpoints for latency, an MCP server for agent ecosystems, and an open schema for everything else.
Self-serve keys
Generous free tier, public docs, OpenAPI spec. Prototype without talking to sales.
Low-token responses
Structured JSON with size, confidence, per-zone fit, and alternative. Built for conversational checkout.
Auditable answers
Every response carries engine version and chart provenance, so agent recommendations can be replayed and verified.
Open by design
Published schema, llms.txt, machine-readable methodology. The citable source for sizing questions.
Let's talk about what sizing could be doing for your store. Talk to sales →
When an AI agent asks
"what size should they buy?"
SizeGuide answers in structured, explainable JSON. Fast enough for conversational checkout, auditable enough to trust.
One conversation, two views: the shopper's chat on the left and the exact API traffic behind it on the right.
Stateless fast path
A single REST call returns the full recommendation. No sessions, no orchestration, edge-served.
MCP server
Native tools for agent ecosystems: recommend_size, convert_size, get_size_chart, explain_fit.
Open sizing schema
A published, versioned JSON standard for charts and fit responses. Free to adopt, built to be cited.
Direct answers, as always
What does SizeGuide need from shoppers?
As little as nothing. Level 1 guidance uses the chart alone; two to four inputs give a quick recommendation; more measurements sharpen it. We never force a long survey, and the confidence score reflects how much we actually know.
Where do your size charts come from?
Three sources, always labeled: official brand data (supplied by the brand), manufacturer-published data (from the brand's own chart, with the source URL kept), or SizeGuide reference data (our synthesis, methodology linked). Every chart is versioned and shows its last-verified date.
What happens when you're not confident?
We say so. Between sizes, you see both options and which fits your preference. Outside a chart's range, the engine declines rather than guessing. A wrong size costs you a return and costs us your trust.
Is this AI? What's actually under the hood?
A deterministic engine: validated charts, measurement logic, and published ease rules. Purchase and fit outcomes tune it only after real evidence accumulates, and any adjustment is disclosed in the explanation. No black boxes, no invented accuracy claims.
What does it cost?
Merchants: free basic charts, $29 for smart charts, $99 for personalized recommendations and analytics, custom for enterprise. Flat tiers with no order-volume surcharges. Consumer tools and the developer free tier stay free.