LLM-powered naming·claude-haiku-4-5

Describe your idea. Get the name.

A real LLM reads your brief, picks the naming style, and proposes domains. Constrain length, force or forbid words, and re-roll until something clicks.

AI Generator · Claude haiku

Describe your business, get brand-ready names.

An LLM reads your brief and proposes domains in the style you choose — checked against the registry in real time.

0 chars
Try:
Auto
5–12 letters
to
optional
optional
apponline
Proposed names

Drop in a brief above to begin

Describe what you're building above, pick a naming style, and hit Generate.
Available — register at standard pricing Premium — marketplace listing Registered — currently owned
How it works

More than a word salad.

A generator that respects your brand. We bias toward short, pronounceable, register-able names — not dictionary mashups.

Word-blend engine

Pulls from 130+ brandable seed words and blends them into novel, sayable names — never just /usr/share/dict.

Live availability

Every generated name is checked against the registry in real time. You only see options you can actually own.

Prefix + suffix control

Bias the output with a theme prefix ("get", "try") or a suffix ("ly", "hq") to fit your existing brand system.

Premium picks surfaced

When a generated name happens to be a premium listing, we flag it inline with the seller's asking price.

Stop second-guessing.

Find your domain in the next 30 seconds.

Free forever. No login required. 1,100+ TLDs in one query — plus AI suggestions and the premium marketplace.

1,147
TLDs
200ms
Avg lookup
100
Suggestions/query
FAQ

AI naming, demystified.

The Generator shuffles a curated word pool — fast, but topic-agnostic. AI reads your brief, infers the right metaphor, and proposes names tailored to the actual product. Use AI when the meaning matters; use Generator when you just want brandable filler.
Each style is a different instruction the model receives. "Brandable" asks for coined, made-up words. "Evocative" leans on real metaphor-laden English. "Two/Three Words" forces clean compounds. "Alt Spelling" creatively respells real words. "Auto" lets the model decide.
Length is enforced both in the prompt and in a post-filter, so anything outside your bounds is dropped before you see it. Exclude is also post-filtered — a banned substring will never appear in your results. Include is a soft preference the model honors when natural.
If the model returns malformed JSON or the request errors, we fall back to the local word-blend generator so the page still has something useful to show. Tap Reroll and you'll usually get the LLM response on the next try.
No — the model samples differently each call, so each Reroll produces a fresh set. Refining the brief (more specifics about audience or tone) usually narrows the field faster than rerolling alone.