Retrograde

A customer-facing storefront to browse menu items, see limited-time drops, place delivery orders, and view order history.

Retrograde Coffee did not have a concrete design for their brand identity and did not have an online presence that represented that brand identity.

A dual-surface full-stack strategy: one product, two interfaces, one shared backend solving both customer ordering and store operations in one cohesive system, with low-friction UX on the customer side and strong control on the admin side.

A single-repo monolith: Next.js UI + Elysia API + Prisma/Postgres, with clear separation between customer storefront and admin operations.

Implemented a full-stack coffee-commerce platform with a guest-friendly storefront and admin control center: menu/catalog management, limited-offer engine (time/capacity + discount rules), cart/checkout and order tracking, cookie-based guest identity with saved delivery data, secure admin auth/workflows, and a typed Next.js + Elysia + Prisma/Postgres architecture for reliable day-to-day store operations.

Balancing a highly stylized retro UI with usability and responsive behavior. Keeping pricing/promo logic consistent between cart preview and server-side order validation. Implementing robust guest-session flows (address persistence + order history) without full auth accounts. Designing admin workflows that are powerful but safe (single active offer, valid order-status transitions). Maintaining strict type/validation alignment across frontend forms, API payloads, and Prisma models.

A single, type-safe full-stack platform that combines a customer storefront and an admin operations dashboard over one shared API and database. On the customer side, the app minimizes friction with guest-based ordering, cart persistence, and straightforward checkout, while still supporting dynamic limited-time offers. On the operations side, admins can manage menu items, promotions, site content, live stats, and order states in one place with guardrails (like validated status transitions and controlled offer activation). This approach replaces fragmented manual workflows with a cohesive system that is easier to run, scale, and maintain.

Operational efficiency and conversion readiness: the app centralizes menu, offers, and order management into one workflow, reduces checkout friction through guest-based ordering (no account required), and enables faster campaign execution via admin-configurable limited drops instead of code changes. While the repository does not include production analytics numbers yet, it is structured to track concrete KPIs such as order completion rate, offer redemption rate, average order value, order processing time by status stage, and repeat guest order frequency.