Face Value

Photo in, price out. Upload a photo of something you own; a vision model identifies it, and the app looks up current eBay listings to estimate an average asking price. Not a sale, not an appraisal — a starting point. Access is restricted to a small, pre-set allowlist of email addresses via passwordless magic-link login.

Stack

• Backend: Go, chi, pgx, sqlc, golang-migrate
• Vision: Hugging Face Inference Providers (OpenAI-compatible router)
• Pricing: eBay Browse API, behind a pricing.Source interface
• Images: Amazon S3 (private bucket, presigned URLs)
• Frontend: Nuxt 4, Tailwind CSS v4, Pinia
• Database: PostgreSQL 16
• Local dev: Docker Compose (Postgres, Mailpit, MinIO)

Project layout

backend/ Go API server (cmd/server, cmd/migrate, internal/...)
frontend/ Nuxt 4 SPA
scripts/ Forge deploy script and API daemon wrapper
docs/ Deployment, eBay, and Hugging Face setup

Deployment

See docs/DEPLOY.md for production on a VPS with Laravel Forge (push-to-deploy from GitHub), and docs/EBAY_SETUP.md / docs/HUGGINGFACE_SETUP.md for the two external API credentials it needs.

Local development

1. Copy .env.example to .env and fill in the values (Hugging Face token, eBay sandbox keys, allowed emails, etc.).

2. Start everything with Docker Compose:

   docker compose up --build

   • Backend API: http://localhost:8080 (/healthz for a liveness check)
   • Frontend: http://localhost:3000
   • Mailpit (magic-link emails): http://localhost:8025
   • MinIO console: http://localhost:9001
   • Postgres: localhost:5432

Backend only

cd backend
go run ./cmd/server
go test ./...
go run github.com/golangci/golangci-lint/v2/cmd/golangci-lint@latest run ./...

Frontend only

cd frontend
npm install
npm run dev
npm run lint
npm run test
npm run build