Why we're building Selify
Merchants with large catalogs are running them with tools that either can't see the catalog or can't talk to a shopper. We're building the layer that does both — carefully, and in the open about what's ready.
Origin
Selify started in 2025 inside Up Go Corp., a Toronto product studio, when we tried to run listings and DMs for a store with AI tools that couldn't see the catalog. The bottleneck wasn't the model; it was context and trust — the AI had to know the products, and the merchant had to be able to check every claim it made. That became the record of every product fact and where it came from, the approval console, and everything around them.
Principles
Outcomes over seats
We charge for work the AI finished, never for logins. Adding a teammate never adds a line item.
Facts before prose
A listing is only as honest as its facts. Every fact has a source and a status, regulated ones come only from you, and copy is written from those facts rather than typed over them.
Nothing silent
Every action is a proposal you can see, approve and undo. The agent holds what it isn't sure about instead of guessing at a shopper.
Your workspace is yours
Isolation is enforced in the database, not just in application code. We don't train models on your data, and inference runs under our own accounts.
Team
Selify is built by a small team at Up Go Corp. in Toronto, working with AI coding agents; we'd rather show you the product than a headcount.
hiring: hello@selify.ai
What's next
- in progress photo-reading vocabulary beyond apparel · CSV upload in the dashboard · Etsy (awaiting Etsy approval) · Amazon (awaiting production access)
- later TikTok · SOC 2 · EU data residency
early access: live in production for a small number of workspaces. the first merchants with large catalogs are onboarding now.