Resonate27

Most personality products start by asking users to repeat information their everyday AI assistant may already know. Resonate27 starts with that existing relationship, without asking users to upload their chat history.

Users copy a prompt into the AI assistant they already use and paste back a compact RESONATE27=[...] vector. Resonate27 turns it into a 27-part personality profile for self-discovery and compatibility-based matching. A 54-question fallback is available, and the first experience works without an account.

Beyond self-discovery, Resonate27 is designed to support personality-based friendship and dating matching. The goal is to help people connect through patterns in how they think, communicate, and relate, rather than relying only on surface-level profiles or swipes.

Resonate27 existed before July 13. During the submission period, I used Codex with GPT-5.6 to add the guest PX27 summon flow, source-aware local guest persistence, an accessible reveal-to-auth handoff, and the 27-dimension detail experience. The qualifying commits begin after the official submission-period start and are listed in the README.

I directed the product and UX decisions. Codex with GPT-5.6 helped implement the Next.js and TypeScript changes, write regression tests, examine local-storage and focus-management edge cases, and verify production deployment.

The main challenge was letting signed-out users see value before login while keeping guest data explicit, local, validated, and recoverable after authentication. The reveal supports skip, Escape, reduced-motion preferences, and reliable focus transfer.

Built With

• codex
• gpt-5.6
• next.js
• postgresql
• prisma
• react
• typescript