In the lab · Work in progress
Picky.
Find the one worth buying.
Tell Picky what you’re looking for. It brings listings together, weighs price, condition, and seller feedback, then explains its strongest pick and a worthwhile alternative.
01 / 05
Search in motion
Watch listings arrive as Picky works through the options.
Swipe to explore more screens
01 / The idea
More listings shouldn't mean more doubt.
A search can turn up dozens of similar products, each with a different price, condition, and seller story. Comparing them across tabs takes time, and the cheapest listing isn't always the one you'd feel good about buying.
Picky puts the options in one place, then looks beyond the headline price. It weighs the listing details and seller feedback, shows how the candidates compare, and explains why one stands out. You also get an alternative when there’s another option worth a look.
Jiji is the first connected marketplace. The product is still in the lab, with Jumia, Konga, and Amazon shown as future sources rather than available searches today.
02 / The experience
From one search to a considered choice.
Say what you want.
Search for a product in plain language. Picky starts with live Jiji listings and gathers the details behind each result.
See the options unfold.
Listings arrive as they’re found. Picky compares condition, price, specifications, and seller feedback in one view.
Choose with context.
Get a recommended listing, the reasons it won, and an alternative worth considering. Open the original listing when you’re ready.
03 / Under the hood
A live search with a reasoned finish.
I built Picky across the frontend, backend, and AI workflow: from a focused search interface to live product updates and a recommendation that explains itself.
Marketplace data, made useful
The backend retrieves Jiji listings, product details, and available seller reviews, then presents them in a consistent format for comparison.
Progress you can actually see
Server-sent events stream status updates and listings to the interface before the comparison and final recommendation arrive.
A recommendation with reasons
AI workflows compare the candidates and return a leading pick, the thinking behind it, and an alternative. The marketplace layer leaves room for more sources as they’re built.
04 / The toolkit
Built with the right tools.
A selected look at the technologies behind the experience.
