Day 87: Deploying the full Measurement Engine endpoint
The single-call engine, live
Deploy the complete engine as the one-call endpoint from Day 75's contract: image + height in → {measurements_cm, size, confidence, warnings} out, running the full four-model pipeline with calibration, error handling, and confidence. Front it with NestJS, call it from a Next.js page, and you've met the core Stage 2 exit criterion — a deployed pipeline that turns a photo into a defensible size recommendation.
from fastapi import FastAPI, UploadFile, Form
import cv2, numpy as np
app = FastAPI(title="FitXpert Measurement Engine")
@app.post("/measure")
async def measure(file: UploadFile, height_cm: float = Form(...)):
img = cv2.imdecode(np.frombuffer(await file.read(), np.uint8), cv2.IMREAD_COLOR)
result = measurement_pipeline(img, real_height_cm=height_cm)
return result # {measurements, size, size_confidence, warnings}Build the demo in the Schemax design system
The roadmap suggests rendering the demo UI in the house design system (navy/teal/amber) from Stage 2 onward — every milestone doubles as a slide, and this measurement demo can sit beside the MES and Supplier Portal kits. 'AI-powered fit' is a credible product line downstream of Xpparel; one budget, two outcomes. A polished demo now is sales collateral later.
Key terms
- Measurement Engine
- The complete deployed service turning a photo + height into calibrated measurements, a size, and a confidence.
- Single-call API
- One endpoint encapsulating the whole multi-model pipeline behind a clean request/response.
Why does the roadmap suggest building the FitXpert demo UI in the Schemax house design system from Stage 2 onward?
