Inspiration

Community clothing exchanges can help people reuse garments, but an item listed only as a size can be hard to judge before a pickup. FitRelay explores a small, practical way to make self-measurement useful: compare a garment you already like with donated garments before visiting an exchange. It is designed with budget-conscious exchanges and people for whom an unsuccessful trip is inconvenient in mind. The project addresses Code for Humanity's measurement-related waste prompt through a clothing reuse use case.

What it does

FitRelay is a free browser app that ranks donated garments by how close three laid-flat measurements are to a reference garment: chest width, shoulder-to-hem length, and shoulder-to-cuff sleeve length. People can add, edit, and remove up to 30 garments, see the signed measurement differences, and shortlist items for in-person inspection.

The initial three garments and all their measurements are fictional demonstration data. The app does not predict how clothing will fit a person. Fabric, cut, condition, and personal preference still matter. No exchange has piloted it, and no reduction in textile waste, travel, or returns has been measured.

How we built it

The app uses accessible HTML, responsive CSS, and native JavaScript modules. The comparison engine validates 1–300 cm inputs, calculates a weighted normalized absolute difference across the three dimensions, and sorts with a stable tie break. The weights and tolerances are explicit design defaults, not a trained or scientifically validated model. No account, backend, analytics, or measurement upload is used; entries live only in the current tab.

The prototype was developed with Codex during the hackathon period. Codex implemented the app and performed the recorded engine and browser checks. The existing engine checks cover ranking, tie breaks, and invalid measurements (3/3 passing); the recorded desktop and mobile browser checks included adding and removing an item. The 28-second narrated video shows the real app and its fictional sample.

Design lesson and next steps

Showing the actual centimeter differences is more honest than presenting a fit percentage. The next useful step would be a small opt-in pilot with a community exchange, with consent-based feedback on whether the tool helps people shortlist garments. Measurement instructions and accessibility should also be tested with real users before claiming any impact.

Try it and inspect the source

Live app: https://estona815.github.io/fitrelay-code-for-humanity-2026/
Public source: https://github.com/estona815/fitrelay-code-for-humanity-2026
Source ZIP: https://github.com/estona815/fitrelay-code-for-humanity-2026/releases/download/demo-v1/code-for-humanity__fitrelay-source__2026-09-27__v001.zip
Demo video (28 seconds): https://youtu.be/-Gd3BFRfz5Q

Run locally with python3 -m http.server 8000 from the source directory and open http://localhost:8000. Run the engine checks with node --test test/*.test.mjs.

AI and third-party disclosure

Development and submission preparation used Codex for implementation, write-up, and test/browser checks, OpenAI ImageGen for the fictional garment art and initial visual concept, and the system text-to-speech voice for narration. The app uses no third-party clothing catalog and no proprietary pre-existing code. The US EPA textile data is background context only; the EPA has not evaluated FitRelay.