Methodology
We hunt the internet.
The write-up stays honest.
Every search and every guide runs the same pipeline. Here it is, end to end — including how the money works.
Read the product
You paste a link (or a photo). We fetch that one page, read its structured data, and build a profile: category, silhouette, materials, colors, and the listed price.
IN A product link → OUT A product profile
Write the hunt
Claude turns the profile into a handful of shopping queries a sharp human would type — style words, not brand words.
IN A product profile → OUT 3–5 search queries
Sweep the stores
The queries run against live shopping data across the web. We merge, dedupe, drop the original brand's own listings, and drop anything that isn't meaningfully cheaper.
IN 3–5 queries → OUT A candidate pool
Rank the twins
Every candidate is scored 0–100 against the profile — silhouette and category first, then materials, colors, and style. Accessories and lookalike-adjacent noise get pushed out.
IN A candidate pool → OUT Ranked twins with scores
Write it up honestly
For guides, the ranked set is annotated — named criteria, honest caveats per pick, and an anti-pick when something deserves it. A human reads every guide before it ships.
IN Ranked twins → OUT A published guide
What we will not do
- Claim a twin is the original. Results are visually similar alternatives, not identical products — that sentence is on every results page for a reason.
- Use a brand’s photography as our own. Product images appear only as attributed search results linking to their store.
- Let a store pay its way up the ranking. Twins are ranked before monetization is considered.
How the money works
When you buy through a Twinz link, the store may pay us a commission at no extra cost to you. Ranking happens first, monetization after — a store never scores higher because it pays more. Stores we have a direct partnership with carry a small check mark.
See it work on something you actually want.
Paste a product link →