Embeddr

Dating for AI agents. Find your nearest neighbor.

A workshop app for runtime control with Unleash and an AI coding assistant.

The frontend (Vite, React, Tailwind CSS v4) renders whatever the backend returns. The backend (Fastify) evaluates every feature flag server-side with the Unleash SDK, serves Auto-Rizz opening lines behind the auto-rizz flag, and reports Ick taps to Unleash as an impact metric. Turning a flag off stops real server-side work, and the running app reflects it within seconds without a redeploy.

Run it

Prerequisites: Node.js 22 or later, and an Unleash instance with a Backend token.

1. Copy server/.env.example to server/.env and fill in the Unleash connection values. The frontend needs no configuration for local use; .env.example documents VITE_BACKEND_URL for deployed setups.
2. Run npm install. This also installs the backend dependencies.
3. Run npm run dev.

The frontend runs at http://localhost:5173 and the backend at http://localhost:3001. Use npm run dev:web or npm run dev:api to run either side alone. Without Unleash credentials the backend still boots, every flag evaluates to off, and the app renders the plain Embeddr experience.

Deploy

The repo deploys as one Railway service: railway.json builds the frontend and starts the backend, which serves the built frontend and the API from a single port. Set UNLEASH_API_URL, UNLEASH_CLIENT_TOKEN, and METRICS_INTERVAL in the service variables; Railway provides PORT.

The pieces

Path	What it is
src/	The frontend. No Unleash SDK, no flag logic, no token.
server/	The backend. Flag evaluation, openers, the ick impact metric. See server/README.md.
WORKBOOK.md	The attendee workbook. Start here at the workshop.
FALLBACKS.md	Plain commands for every assistant-driven step.
CLAUDE.md, AGENTS.md	Guidance for AI assistants working in this repo.