VS Fit QC Plugin

VS Fit QC captures tension, pressure, and textured renders for all views of the garment, scores each view against a category threshold profile (casualwear, activewear, denim, etc.), and produces a multi-page PDF report. AI Analysis is off by default — the core scoring/PDF pipeline never contacts the network unless you turn it on. With it enabled, six specialized Claude agents add a written fit report: garment identification, fit standards citation, overall fit summary, key areas of concern, pattern-data analysis, recommendations, and a final prioritized adjustment schedule.

This is the default flow: scoring and the PDF report, no network calls.

Open VS Fit QC against an open garment and click Run Fit QC. It captures tension, pressure, and textured renders for all views and scores each against the category profile.

Every run produces a timestamped folder in ~/Downloads/Plugin Output/VSFitQC/<GarmentName>_<timestamp>/:

Reports are safe to zip and share.

Adds a written, multi-agent fit report on top of the same capture/scoring pipeline. Requires your own Anthropic API key and is off until you explicitly turn it on.

Click the AI Analysis toggle (top of the panel, off by default). The first time, a consent overlay explains exactly what gets sent to Anthropic — rendered images and pattern geometry — and asks you to confirm before anything is transmitted.

Pick a depth for the AI pass:

Click Run Fit QC as usual — the capture and scoring pipeline runs the same way, with the six-agent AI analysis added on top.

Same output folder and PDF as the non-AI flow, now with the written fit analysis (standards citation, key areas of concern, recommendations, prioritized adjustment schedule) included.

An AI run that fails silently should be checked against the debug log (captures every AI agent call with model, token budget, and outcome) at Browzwear/VStitcher/Logs/vs-fit-qc/vsfitqc_ai_debug.log.