Metr Fit for developers

What the API does

You send your product catalog, available variants and approved size charts. Your backend requests a questionnaire, submits customer answers and receives a size recommendation with confidence, fit details and an explanation. All customer-facing requests go through your own backend.

• One merchant can own multiple stores. Every resource is isolated by merchant and store.
• Sizing runs in three modules: every chart is normalized to body centimetres, explicit answers (height, weight, tape measurements) become body estimates with stated uncertainty, and implicit answers (usual size, a size worn at another brand, body-shape illustrations) refine them. When the operator configures it, TypeSafe's Jev model then chooses among the plausible, in-stock chart sizes. Nothing can invent a size or a body measurement.
• Adult tops, bottoms, dresses/one-piece garments and supported shoe charts are covered. Unsupported products return an explicit error.
• Confidence is an uncalibrated evidence score, not a probability of fit or a measured accuracy claim.

Start here

Follow the Quickstart, then read Authentication and Integration flow before putting Fit on a product page. The API Reference below is generated from the Go route registry and wire types, not a proposed SDK.

export METR_API_URL=http://localhost:8080
# Run the API repository locally first. Use HTTPS for a deployed service.
python3 scripts/demo.py

Version and deployment

All product endpoints use /v1. Deploy your own API from metr_api; this documentation does not imply a shared hosted endpoint. Health checks are /healthz and /readyz. The OpenAPI contract is available at /v1/openapi.json on your deployment.