Web Applications / Applied AI / E-commerce / Mobile
VinoSeeker: a wine discovery platform for drinkers and wineries
A wine discovery app for drinkers and wineries, with tiered Stripe subscriptions, an AI sommelier grounded in its own wine data, and label scanning from a phone camera. It runs on the same AWS stack we build for clients.
- In development and live
- Since 2021
- Wineries, vineyards and tasting rooms
- 4,000+
- Wines in the catalog
- 60,000+
Overview
An independent business we've built from the start.
VinoSeeker helps people find wineries, keep track of what they taste and manage their cellar, and gives wineries a listing they can claim and improve. Event organizers run multi-winery passport events on it.
It's an independent company, founded by SimBco's founder, and SimBco has built it since the beginning. It grew out of a gap our founder ran into as a wine lover: no tool kept track of the wineries you'd been to and the ones you wanted to visit. Oregon's Willamette Valley alone has hundreds of wineries. Some have national distribution, but many are family farms that depend on people coming through their tasting rooms.
Because we're this close to it, VinoSeeker is where we prove new frameworks and methods before we recommend them to clients. Interns from a local code school have also gained experience on it, working in a large, real production codebase. It has grown from user feedback over several years into a full tool for wine lovers and the wine-curious.
The Challenge
Getting the model of the wine world right.
On client projects we interview the people who know the field. Here we started as wine drinkers, not industry insiders. The technology was the simple part. Modeling the data correctly was the hard part.
People lump wineries, vineyards and tasting rooms together, but they do different things. A winery is a business that makes wine, so it needs its wines and vintages tracked, with tasting notes year over year. A vineyard grows grapes; it has an AVA and a soil type, and it's where the farming happens. And you don't visit either one unless it's also a tasting room, where you can taste and buy.
Our Solution
A flexible model, subscriptions, AI and a PWA on AWS.
A place can be any combination of winery, vineyard and tasting room, each with the information that role needs, and the model is built so new kinds of place, such as wine bars, wine shops and tour companies, can be added.
Five consumer tiers, from free to Connoisseur, each unlock features on top of the last: unlimited library, label scanning, then progressively more capable AI. Payment runs through Stripe Checkout and the Stripe billing portal, so card data never touches our servers. Upgrades take effect immediately with proration; downgrades wait for the end of the period. Signed webhooks keep every subscription in sync, and the API enforces each tier's features on the server, not just in the app.
Ava, the AI sommelier, answers from VinoSeeker's own catalog: listings and wines are embedded into a vector index in Postgres, and Ava calls tools to search wines, find nearby wineries and work with the user's cellar. Each tier uses the model that fits it, so higher tiers get more capable models and lower tiers stay affordable to serve.
Point the camera at a label or a tasting-room menu and a vision model reads it into structured fields, which we match against the catalog. The app is a progressive web app with push notifications, so there's no app store in the way, and it runs on AWS defined in code: API, database, image pipeline, scheduled jobs and a Next.js marketing site on OpenNext.
The Results
For wine lovers, and for the wineries they visit.
Users track where they've been, what they drank and what's in their cellar, and share what they love with their circle of friends.
Businesses claim their listings and keep their information correct. More than that, they can stay in touch with the people who visit, and help guide someone from a first tasting toward becoming a regular, and ideally a wine club member.
For us, it's proof that the patterns we recommend hold up in production. The hosting setup behind this site came from VinoSeeker.
Technology used
- React
- Ionic
- Feathers
- PostgreSQL
- pgvector
- AWS Bedrock
- Stripe
- Next.js
- AWS CDK