Names, with the meanings labelled

• Syllables, by a documented English spelling heuristic with a short list of exceptions. It is a free in-browser syllable counter, and every syllable number the shortlist claims is checked back against it.
• Initials and the engraved monogram, computed separately, because they are two different strings that spell different things: initials run first-middle-last, an engraving runs first-last-middle. A name can be clean spoken and awkward on a towel.
• Surname clash detection — full rhyme, half rhyme, alliteration, a seam collision where one name ends on the sound the next begins with, a vowel elision that slurs the join, and names that read equally well in either direction. All of the collisions that come out of how the two names rhyme or run together.
• Nickname derivation, split into short forms that fall out of the spelling by rule and the irregular conventional English diminutives that no rule produces.
• Spelling risk — where English readers will say a name more than one way.
• A sibling check: the syllable counts and endings of names already in the family are measured here and handed over as facts, so "harmonises with" means something specific.

Everything computable from the letters alone is derived here rather than asked for, and it costs nothing.

A name's meaning and origin are factual claims a parent cannot check, made once, about a decision that is permanently in place. A model asked to describe an etymology will produce a confident, plausible, wrong one and nothing downstream catches it. So every meaning carries an attestation label — well-attested and settled, disputed between specialists, a folk etymology everyone repeats, a modern coinage with no ancient meaning, or genuinely uncertain. Anything that is not settled has a note explaining what the competing readings are, and that note is required: a missing label is a hard finding, not a blank field.

No popularity rankings anywhere. Familiarity comes back as a band and a direction, never a chart position. Popular-name rankings are re-cut every year and go stale within one, and a stale number reads exactly as confident as a fresh one. A rank appearing anywhere in the reply is flagged automatically.

Pronunciation comes back as an English respelling, and the page counts its beats against the syllable count. The downside of each name is a required field, because it is the one that makes the rest worth reading.

Every claim in the answer is then re-derived from the letters and reconciled against what this page computed. Each candidate that disagrees carries the disagreement next to it: a wrong syllable count is displayed, never silently rendered. Nothing here is stated as settled fact that the app cannot show its working for.

A refinement is a second pass against the same brief rather than a fresh run. It holds the brief steady and changes only what you asked to change: anything you say to keep comes back with the same meaning and the same attestation label, and what you did not mention stays put.

Shortlists are saved to your SkillSafe account when you are signed in, and mirrored in this browser for an instant paint. There is semantic search across them, over both the names and the brief that produced them, so "the Welsh one with the sea meanings" finds it.

Exports: a plain-text shortlist, Markdown, JSON, a candidates CSV, and print. The names alone copy separately from the whole answer.