Naming my side project took 2 weeks until I treated it like a search problem

I spent two weeks naming a side project and produced exactly four candidates, all bad. The stall wasn't creativity — it was that open-ended naming has no stopping rule. Every name I thought of spawned three more directions, and none felt finished. I needed constraints, not inspiration.

So I reframed it as a search problem: given an industry and a vibe, return a fixed set of candidates I could react to. That single change — bounding the output — broke the loop.

"Vibe" is the missing input

The reason my brainstorm failed is that I never stated what the name should feel like. Playful? Trustworthy? Technical? A name that's perfect for one vibe is wrong for another, and without picking, every option looked equally plausible.

When I forced myself to choose a vibe first, the candidates got sharper. "Calm and competent" ruled out the pun names I kept drifting toward. BrandName Gen takes an industry and a vibe and returns ten on-brand name ideas you can riff on — which is the bounded list I was trying and failing to produce in my notes app.

Ten is the right number to react to

Too few candidates and you pick from scarcity; too many and you're back to analysis paralysis. Ten hits a sweet spot: enough to see patterns, few enough to actually read. I learned to scan for the two or three that share a feeling I like, then iterate on those rather than starting over. Anything past that and the list starts to blur into noise I scroll past.

The copy-list output matters more than it sounds. I wanted the names as plain text I could paste into a trademark search, a domain checker, and a group chat — not locked in a tool's gallery. BrandName Gen returns a copyable list, which fits how naming actually proceeds: generate, then verify availability elsewhere.

The honest caveat

A name generator doesn't check trademarks or domains. Its job ends at "here are ten directions." The availability work is still yours, and it should be — that's where the real constraints live. Treat the output as a prompt for your own judgement, not a finished brand.

Why I stopped asking friends for names

The worst part of manual naming was the group chat. I'd post three candidates and get twelve more, none evaluated against the vibe, all now carrying someone's invested opinion. The conversation drifted from "which fits" to "which does Dave like," and I'd ship a name nobody hated instead of one I wanted.

Generating the list myself, bounded and vibe-tagged, cut that loop out. I could react to ten candidates in private, pick the two with the right feeling, and only then bring specifics to anyone else — by which point I had a recommendation, not a question. The tool didn't make the name; it removed the social friction that had been the actual blocker.

FAQ

Why ten names and not three or fifty?
Ten is enough to see a pattern in what you gravitate toward, but few enough to read without fatigue. Three feels like scarcity; fifty recreates the paralysis.

Does picking a vibe limit good accidents?
It focuses the list. You can always run a second pass with a different vibe if the first felt off — that's cheaper than open-ended brainstorming.

Will it check if the name is taken?
No. It generates candidates. Domain and trademark availability are separate steps you run yourself.

Is this better than a thesaurus?
A thesaurus gives words; this gives names tuned to an industry and a feeling. Use it when you need direction, not just synonyms.