Gift Squire

Inspiration

A gift can be thoughtful and still feel unlike the person receiving it.

Choosing clothing for someone means balancing two wishes: you want a lovely surprise, and they want something they would enjoy wearing. A product photograph alone cannot give both people a shared way to explore that choice.

That inspired GiftSquire: give them a say, keep a little surprise.

We wanted a warmer decision process for birthdays, celebrations, festive occasions and everyday generosity. Shopping for yourself belongs in the same experience: seeing the possibilities should help you express your preferences and make a considered choice.

What it does

GiftSquire creates a private space for exploring outfits for yourself or someone else.

A giver records the occasion, budget and style wishes, then builds a shortlist. They can invite the recipient to join privately and provide their own photo, or prepare proposals from a photo the recipient explicitly permitted them to use.

YouCam powers the visual experience. Its clothing virtual try-on turns the reference outfit and permitted photo into personalised previews. Optional YouCam lipstick try-on adds a separate close-up alongside the full outfit, with saved shade comparisons.

The recipient reviews and approves the options they would welcome. The giver makes a private final choice from those approved looks, preserving a little surprise. Approvals belong to the current preview: replacing a photo or recreating an outfit clears the affected approval so the choice stays connected to what was reviewed.

Users can complete setup through User Mode or talk it through in Interactive Mode. They can refine overall wishes, an individual outfit or a saved decision using Chat or Discuss with AI. Changes are presented for review and explicit confirmation before saving.

The outcome is useful beyond the screen. A private Partner Brief keeps the shortlist, planning context and image revisions together. A separate Gift Brief presents selected preview images and a personal message, with review before email delivery. Users can revisit saved briefs, open a printable/PDF view and download actual preview images.

How we built it

GiftSquire uses Next.js, React and TypeScript for the web experience, Convex for server-side data and workflow, and Vercel for hosting through its GitHub integration.

YouCam has an essential job in the decision loop: it creates the clothing and optional lipstick previews that people inspect before approving an option. The original outfit image, source-photo revision, generated result and approval state remain connected in the application.

OpenAI powers the Style Guide: GPT-6 Luna helps interpret wishes and prepare structured proposals, while GPT-Live-1 supports live conversation. The application controls permissions, approved-choice checks and confirmed saves. A conversational response cannot substitute for a generated preview or a recipient's approval.

Convex Auth and Resend support verified email/password accounts and recovery. Resend also delivers reviewed Gift Briefs. Provider credentials stay server-side. Sharp supports image processing, while schemas and version checks keep requests and saved records consistent.

Challenges we ran into

The most important challenge was representing both people honestly. A giver can prepare a proposal, but only the recipient can approve a gift. A shared preview also must not silently expose the recipient's original uploaded photo.

We made that distinction part of the workflow through separate roles, explicit preview sharing, revocable invitations and version-bound approvals. Saved choices and briefs retain the context of the images actually reviewed.

Live voice presented a different challenge: connection acceptance is only the beginning. The user needs usable audio, captions, a clear review and a confirmed saved result. We added proactive welcomes, editable reviews, retained conversation drafts and deliberate reconnect/retry options. An interrupted conversation can still lead to reviewing the captured material.

We also worked to make a multi-step process feel approachable. Numbered sections, a speech bubble tied to the active mode, visible status, four appearances and restrained glass effects help users understand where they are and what comes next.

Accomplishments that we're proud of

We built and exercised a complete hosted journey: create a shopping space, involve a separate recipient, generate real YouCam previews, record approvals, preserve a private final choice and create useful saved briefs.

Production acceptance included a real clothing output, lipstick close-up, recipient-only approval, Partner Brief/PDF, image download and a Gift Brief reported delivered by Resend. A later recorded walkthrough generated two outfit previews and a lipstick close-up, captured both recipient approvals and followed the reviewed email-sharing flow.

Both live voice modes completed real provider conversations and reviewed handoffs. Outfit-specific spoken feedback saved to the intended record while preserving the existing preview. The application also supports written feedback, confirmed context changes and recovery paths.

The project combines a distinctive two-person decision with practical product detail: self-shopping, clear permissions, account verification/recovery, accessible interaction alternatives, saved output and a coherent visual identity. Its approved English-captioned walkthrough runs 2 minutes 42 seconds in 1080p.

What we learned

Virtual try-on becomes more useful when the image leads to a decision people can participate in and revisit.

We learned to give each part a clear responsibility: YouCam creates the personalised visual possibilities; the Style Guide helps people express and refine wishes; the application preserves permission, review and approval. The recipient's own decision gives the gift its personal meaning.

We also learned that recovery belongs in the product experience. Keeping a draft, explaining what happened and offering a clear next action helps a person continue without repeating everything or losing control of a change.

What's next for GiftSquire

We want to extend merchant catalog connections, add reliable product/pricing information, explore purchasing integrations and refine the experience through direct giver/recipient feedback. We will continue improving mobile and voice interaction while preserving the simple private-space workflow.

The goal remains the same: make choosing for someone feel personal, collaborative and a little surprising.

Thank you, YouCam and Perfect Corp., for the previews at the heart of GiftSquire's experience.

GiftSquire. Choose with confidence. Give with thought.

Built With

• clothing-virtual-try-on
• convex
• convex-auth
• css
• gpt-6-luna
• gpt-live-1
• makeup-virtual-try-on
• next.js
• node.js
• openai
• playwright
• react
• resend
• sharp
• typescript
• vercel
• vitest
• webrtc
• youcam-api
• zod