Inspiration

A journey through Amman, Petra and Wadi Rum brings different clothing decisions: a city day, a walking day, and an evening at camp. Packing often means guessing what goes together and bringing extra pieces just in case.

Rihla starts with the traveler's wardrobe and a specific route. It connects an appearance preview to a practical decision: what belongs in the bag?

What it does

Travelers choose a stop, select multiple garments, upload a person photo, and consent to a YouCam preview. Rihla applies supported clothing sequentially, using each generated result as the input for the next garment. The traveler can compare the original photo with the final result and add the selected pieces to the destination's packing plan.

A single checklist gathers garments across all three stops. Reused pieces appear once, with their destinations attached. Travelers can check items, download a text plan, and print or save a PDF. The mobile-first flow uses focused screens, large selection cards, a selected-piece tray, and sticky next actions.

The starter garment photos are clearly labeled AI-generated samples. Travelers can add their own clothes. Scarves and other accessories can be selected for packing; they are not rendered in the clothing preview.

How we built it

Rihla uses React, TypeScript, Vinext, Vite and Cloudflare Workers. The server integrates YouCam AI Clothes Virtual Try-On V4 through image registration, signed uploads, asynchronous task creation and polling.

Multi-garment preview is our orchestration of separate API tasks, not a claim that one request accepts several garment images. A signed ticket identifies the completed previous task; the server verifies its success and passes its output URL to the next task. Delayed requests can be checked again without recreating the active task. Earlier successful steps are retained during the session if a later step needs retrying.

The API key stays on the server. Photos and generated results remain in browser session memory; saved local plan data excludes them. Submitted images are also subject to YouCam's processing policies.

Challenges we ran into

The first challenge was making try-on useful beyond the image. Connecting selected garments to destination planning and a deduplicated checklist gives the preview an actionable outcome.

The second was sequential layering: categories need an order, in-flight tasks need recovery, and later generations can change previous details. We show generation counts and limitations before processing rather than promising an exact outfit simulation.

We also fixed a hosted-runtime incompatibility that initially prevented image uploads and redesigned a long desktop-oriented screen into a guided mobile journey.

Accomplishments

We built and deployed the complete route-to-outfit-to-packing flow. The owner successfully used a personal-photo try-on. A separate controlled API test applied a linen shirt followed by a jacket and produced a final image visibly containing both. Twenty-one automated tests cover planning, multi-selection, signed tickets, provider contracts, chaining and retry behavior; type checking and production builds pass.

What we learned

A preview becomes more useful when it helps someone decide what to take. Clear destination context, garment reuse, and a concrete checklist give the generated image a purpose. Equally, a multi-step AI workflow needs honest progress, recovery and limits.

What's next

Test more garment combinations and permitted photos, observe real travelers completing the flow, and expand itineraries. Scarf try-on needs separate integration and fidelity testing. Weather-aware planning and retail integrations remain future work.

Built with

React · TypeScript · Vinext · Vite · Cloudflare Workers · YouCam AI Clothes Virtual Try-On V4 · CSS · SVG · browser local storage · ChatGPT Images

Try Rihla

Public app · Source code and setup

The app is public. Signed-out mobile testing was confirmed on 31 October.

AI assistance disclosure

AI tools assisted with ideation, storyboards, code, tests, writing and interface iteration. ChatGPT Images created the labeled sample garment photos. Actual appearance previews are generated by the integrated YouCam API. Concept storyboards are not presented as functioning app footage.

Built With

Share this project:

Updates

Submission history