How this works
A catalog that shows its work.
Most rankings are bought. Ad budgets buy shelf space, affiliate fees buy "best of" lists, and the price of the thing in your cart is — more often than you'd like — the price of its marketing.
Krekib is a catalog with one rule: products earn their place through evidence. Teardowns, long-term reviews, materials, engineering. No brand pays to be here. No brand can.
The research is done by machines, and we're not shy about it. Agents read the teardowns, the forums, and the test labs; score every product against the criteria that actually matter for its type; and refresh every ranking twice a year. Old editions stay public, so you can check our work.
Every claim carries its receipt. Every ranking shows its criteria — and if you disagree with our weights, move the sliders and make it yours.
We buy nothing, sell nothing, and take nothing from brands.
How the agents work
Each product type gets its own research run. A Scout finds candidates worth considering. A Criteria Analyst asks experts, teardowns, and enthusiast communities what actually separates excellent from mediocre — then locks those criteria and weights into a public document.
The Curator scores every product against those criteria, writes the short verdicts, and ranks the list. A Critic rejects anything that lacks receipts: every product needs at least two evidence URLs, and at least one must be independent of the manufacturer. Missing evidence means the product is dropped — never padded.
An Iconographer builds the covers. Source photos are rehosted with attribution; studio covers are generated in-house so every product has a consistent plate on the shelf.
Scoring math
Rankings are arithmetic, not vibes. For each product we compute:
substance = weighted mean of criteria scores
value = substance + design/2 − hypePenalty
Ranked by value · scores on a 0–10 scale
Substance comes from the type's criteria and house weights. Design is a separate 0–10 assessment of construction and form. Hype penalty (0–3) rises when price sits above the diminishing-returns line without material justification, when the story is mostly brand, or when evidence is thin.
Awards follow the numbers: Top Pick is rank one; Budget is the cheapest product with substance ≥ 7; Upgrade is the dearest product with no hype penalty; Specialist covers a clear niche win.
On every edition page, the criteria sliders start at house weights. Move them and the list re-ranks live in your browser — share the URL and anyone opens your version. That is the trust test: a ranking that survives your scrutiny.
Refresh policy
Editions refresh about twice a year. When a type comes due, agents research again — hunting past the products we already know — and publish a new version. The previous edition is archived, not edited. You can still open it, compare ranks, and read what changed.
Criteria are rebuilt when they are older than twelve months, or missing. Slugs stay stable across editions so a product's page can show its history: where it ranked, and for how long.
Images & takedown
Generated covers are our own works — studio portraits on the same bone ground as the site, so products sit quietly on the page. Real photos are documentary: rehosted from manufacturer press pages when we can, always stored with a source URL and a license note.
If you own a photograph we've rehosted and want it removed, email [email protected] with the page URL and the image. We remove the photo; the generated cover stays.
Independence
No affiliate links. No sponsored slots. No brand partnerships that touch the ranking. Prices are typical street prices as of each edition's date — a snapshot, not a store.
We buy nothing, sell nothing, and take nothing from brands.
Start here
Open an edition. Read a verdict. Hover the evidence. Move a slider.