Skip to content
← All posts

Build story · 22 November 2026 · 6 min read

How We Built ShareMyInvite: A Wedding Invitation SaaS With Next.js, Supabase and Dodo Payments

Inside ShareMyInvite, our wedding invitation SaaS: rupee pricing for a global audience, prices out of the code, a DIY partner programme and WhatsApp video.

By Nagz Studio

Indian weddings are invited on WhatsApp now. The printed card still exists, but the invitation most guests actually open is a forwarded image or PDF on their phone, and most of those look like a scanned card because they are one.

ShareMyInvite is our answer: an animated wedding invitation, personalised in minutes and shared as a single link. Guests tap it and the invitation opens like a card, with the couple's names, the events and the venues and, on higher plans, photos, music and a video version for WhatsApp status.

This post is about how it's built, for developers and founders making something similar. Not the code line by line, but the decisions that mattered and why.

The stack

  • Next.js for the site, the editor and the invitations themselves. Pages are static and cached wherever possible, which suits a product whose traffic can arrive in a burst when a family forwards a link to everyone at once.
  • Supabase for the database and storage: invitations, orders, partners and payouts.
  • Dodo Payments for checkout, in rupees.
  • Resend for email: order confirmations, partner approvals, sign-in links.
  • GSAP and Motion for the animation that makes an invitation feel like opening a card.
  • Vercel for hosting, with its analytics, and PostHog for product analytics.

One choice we'd recommend to anyone: the whole site runs without any keys. Without Supabase, Dodo or Resend set up, the demo invitations load, the editor works, and checkout politely says it isn't configured. A new developer, or a fresh preview deploy, sees a working product in one command, not a wall of errors.

The product in four plans

Each plan is a step up from the one before:

Plan What it adds
Link The full animated invitation, edit anytime, live the moment you pay
Link + Photos & Music A couple photo and gallery, and music
Link + Video Everything above, plus a 9:16 MP4 for WhatsApp status and Instagram
Save the Date Send the date now; the same link becomes the invitation later

The Save the Date is the one we like most as a product decision. Families announce the date months before the invitation is ready. Rather than sell two unconnected things, the same link grows into the full invitation, and the upgrade deducts what they already paid.

Pricing in rupees, for a worldwide audience

The product is made for Indian weddings, but many families have relatives abroad, and sometimes they're the ones paying. We wanted one honest price, not a rupee price for India and a marked-up dollar price for everyone else.

So every plan is priced and charged in rupees. A foreign card still works, and the payment provider converts at checkout. Outside India, the site shows the rupee price with an "about $X" estimate beside it, and says the card will be charged in the customer's own currency.

The interesting part is showing that estimate without making pages dynamic and without a flash of the wrong price:

  1. On the first visit to a page with a price, the edge reads the visitor's country from Vercel's request header and stores it in a cookie for 30 days.
  2. A tiny inline script in the page head turns that cookie into an attribute on the <html> element before the page paints.
  3. The dollar estimate is in the HTML for everyone; CSS hides it for visitors in India.

Pages stay static and cached, nothing flickers, and a visitor we can't place, or one without JavaScript, simply sees the rupee price.

Prices live in one place, never in the copy

Every price lives in environment variables and is read in one file. Content never contains a number. The marketing pages, FAQs and blog posts use placeholders such as {{price:link}}, filled in when the page renders.

That sounds like over-engineering until the first time you change a price. Then it's one setting and a redeploy, and every page, every FAQ answer, every checkout button and the structured data follow. No forgotten blog post still quoting last month's price.

The same rule applies at checkout: the checkout takes a plan, never a price. The browser says "link_photos", and the server picks that plan's product. Nothing a visitor can edit in their browser changes what they pay.

Upgrades without a coupon anyone can reuse

The Save the Date upgrade needed a discount equal to what the family had already paid. The easy way is a standing coupon code. The problem is that a code shown on a checkout page can be copied and shared.

Instead, each upgrade gets its own discount, created at that moment: single use, limited to the invitation products, expiring within a day, and checked against the expected amount before the customer is sent to pay. If the discount can't be created, the checkout stops, rather than quietly charging the full price. We'd rather lose a sale for an hour than overcharge a family.

A partner programme without affiliate software

Wedding planners, photographers, make-up artists and venues talk to couples before they think about invitations, so they're natural partners. Affiliate platforms exist, but they're another subscription and another dashboard for a small product. We built the smallest version that works:

  • Apply: four fields, no bank details. The application lands as a pending row, and we get an email with a one-click Approve button.
  • Track: each partner shares their own link. It counts the click and sets a 60-day cookie; checkout copies the partner onto the order, and the payment webhook records their commission, 20% of every invitation paid for within those 60 days. Refunds earn nothing.
  • Dashboard: a sign-in link by email, valid for 24 hours, opens their link, a WhatsApp share button, a QR code, their taps, paid invitations and earnings.
  • Pay: once a month, by hand. One SQL query lists who's due; after sending the money, one database function records the payout and marks every commission it covers.

Attribution is automatic; only the money moves by hand. At this size, that's the right split. Automating payouts can wait until there's enough volume to justify it.

Video invitations for WhatsApp status

Many families post the invitation as a WhatsApp status, which needs a vertical video, not a link. The Link + Video plan includes a 9:16 MP4 of the invitation, delivered within 72 hours.

The video is made from the same animated invitation the guests see: a script renders it to MP4 with ffmpeg, with the chosen music track. One source of truth for both: change the invitation and the video matches it.

The same idea saves work elsewhere. The sample invitations on the site show a still poster first, and those posters are rendered from the real templates by a script, so they never drift from what a customer gets.

Designs, languages and the details that matter at a wedding

The invitation designs range from South Indian traditional and Kerala to Nikah, church and Italian editorial styles. On the free poster maker, names typed in Tamil or Hindi are set in a font made for that script, and each occasion has Tamil wording to choose from. That isn't decoration. A family that types the bride's name in Tamil and sees it fall back to a generic font won't share it.

Free tools that bring families in

Alongside the paid invitations, ShareMyInvite has a free poster maker: wedding countdowns, engagement, Valaikappu, birthdays, Pongal and more, free to download with no sign-up. Each occasion has its own page, built to be found by the families searching for it. We wrote up what we learned building it as a separate post on free tools as marketing.

What we'd tell another founder

  • Make the product run without its services. Demos, previews and new contributors all benefit.
  • Keep prices out of the copy from day one. Retrofitting it means finding every number on every page.
  • Never trust the browser with money. Plans, not prices; discounts made per order, not shared codes.
  • Build the small version of the thing you'd normally buy, if it's core to growth and the bought version is heavier than your product.
  • Design for how people actually share. For Indian weddings that's WhatsApp, a vertical video and the family's own language.

If you're building a product like this and would rather have it built, that's what Nagz Studio does. And if you're planning a wedding, ShareMyInvite is the result.

Which of these decisions would you have made differently for your own product?