Weddy

The problem

A wedding site is built once, for one day, for one couple — and then it is thrown away while the next couple starts from zero. By hand that is weeks of work per wedding; as a template it means every couple gets the same page. And the invitation is only half of it: the replies and the details have to live somewhere the couple can actually run the day from.

What I built

• One engine, many weddings — a single shared codebase serves every couple, each with their own editable invitation, instead of a copy of the project per wedding.
• It started as a real wedding site that was really used. The engine is a straight port of it, tests included, so the product was built on something proven rather than on a prototype.
• Each wedding is its own tenant — keys and files scoped at the entry point, so the invitation logic itself never had to learn what a tenant is.
• Onboarding takes about two minutes: names, date, a vibe. The new invitation is seeded with the couple's own details from the first second, with nothing borrowed from the reference wedding.
• Editing is passwordless — a signed capability link handed over at sign-up, valid for a week.
• The AI does the parts that wow: a cinematic cover video generated for the day, themes to match, a drafted story, and an animated opening — envelope, curtain or video reveal.

A look inside

The result

Live at tryweddy.com with a founding launch: $20 one-time for the first 50 couples, a lifetime founding price that steps up afterwards. The engine serves a real wedding as its public demo. Payments and the guest-photo engine are the next two steps on the build order.

Built with