Build story · Our product
Peeklyn is our own consumer app. Snap a piece of clothing, pair it with what is already on your rack, and try the outfit on your own photo instead of a model. We designed it, built it and run it, and it is open to everyone.
See it live: peeklyn.com ↗
- Our product
- designed, built and run in house
- Open to everyone
- no invite, no waitlist
- Your own photo
- not a model, not a stand-in
- Our own cloud
- app, database, storage and image models
What it does
A closet you build by taking photos: each piece is cut out of its background and filed. Pair pieces from your own rack, then see the outfit on your own photo before you wear it or buy it.
How it is built
A web app on a container service with a managed database and object storage, its own sign-in, and image models called from the same cloud project.
It started inside a larger platform and was made standalone: its explore feed, sharing, screening and quiz are its own now, with only a visit counter still reporting back.
How we run it
We watch the bill as closely as the uptime. Image work is the expensive part, so cutouts are capped and the service scales to zero when nobody is using it.
What we learned
What we would tell the next team.
- 01Image AI is a cost line before it is a feature. Caps and scale-to-zero were designed in, not added later.
- 02Splitting a product out of a platform is worth doing early: every shared table is a later migration.
- 03Real photos of real people beat a perfect generated model for trust.
Tell me what you're building.
Philip Radvan, founder. Every message lands with me, and it's me on the call.