Score my name. Does it read like a brand?
Type any name. We render its readable wording big — like a logo — score it 0–100 for brandability, and show every way the letters can be read. Same engine that grades ~425,000 dropping .coms every day.
How scoring works
pure textual · no brand knowledgeThe score is a name-quality read, nothing more. It knows no real brands and no popularity — just letters. It looks at two things and keeps whichever scores higher.
Does it split into real words?
We segment the name against a list of the 10,000 most common English words, covering as many letters as possible with as few words as we can. Names that resolve into real words — DeskKeep, AllianceBank — score 65–100. That's also where the CamelCase wording comes from.
Or does it at least read smoothly?
For coined names, we measure letter-pair familiarity — how often each pair of letters shows up in real English. It rescues invented-but-sayable names like awega, but it's capped at 60: from letters alone you can't tell a real brand from junk, so coined names always rank below real words.
What letters can't tell you
The score above stops where letters stop. On every name we confirm is available, a language model then reads it and scores it 0–100 for how good a brand it would actually make — catching what no textual rule can. entrustpilot scores 78 on letters and 30as a brand (“too close to an existing major brand”); coralifeinsurance scores 60 and 30(“misspelling of Life”). That read is the brand column on every list, along with the tags that power browsing by type. This page scores the letters only — paste a name you own and it tells you how it reads, not what a model thinks of it.
The wording is rendered in CamelCase, and when a name reads more than one way we keep every distinct reading — the expertsexchange → ExpertsExchange / ExpertSexChangeeffect. That's the risk (and the fun) of naming: the same letters can be read in ways you didn't intend.