Generation Failed — What Now?

First: if a generation fails for any reason, your treat is automatically returned to your account. You never pay for a failed run. Check your treat count — it should already be back.

Now let's figure out what went wrong.

---

Common Reasons a Generation Fails

1. Your input was too vague

This is the most common cause. The engine needs something to work with. If your idea description is "I'm building an app" or "a startup for business," the model has almost nothing to go on — it generates candidates, but they're too generic to survive the cliché filter, and you end up with zero results.

How to fix it: Be specific about what your product does, who it's for, and what makes it different. You don't need a pitch deck — two or three specific sentences is enough. Compare:

Too vague: "A productivity app"

Much better: "A task manager for freelance designers who work across multiple client projects and need to track time without switching apps"

The more specific your idea, the more targeted the name candidates — and the more likely they are to survive the filters.

2. All candidates were taken

Some niches are brutally competitive in the domain market. Common industry terms, popular two-word compounds, and anything adjacent to finance, health, or crypto tends to have its best .com options long gone.

How to fix it:

Try a different domain extension (.co, .io, .app) and specify that in your generation

Try a different vibe — "Friendly Weird" and "Bold & Direct" tend to produce more creative names that are less likely to be registered

Add more specificity to your idea — unusual combinations of words are more likely to have available domains than common ones

3. API timeout

Rare, but it happens. The engine makes real-time calls to multiple services — the LLM, the phonosemantic scorer, and live RDAP domain checks. Occasionally one of those has a hiccup.

How to fix it: Retry immediately. Timeouts are almost always transient. If it fails three times in a row, drop us a message via the in-app chat.

4. The "Quiet Luxury" vibe with a very specific niche

This one trips people up. Quiet Luxury is the strictest vibe — it filters aggressively for a certain refined, understated quality. When you combine that strictness with a very specific niche (say, "B2B geospatial data for municipal governments"), the overlap between "names that fit this niche" and "names that sound like Quiet Luxury" can be almost nothing.

How to fix it: Try "Surprise Me" — it lets the engine pick the best vibe for your idea, which often breaks through where the strict vibes can't. Alternatively, try "Bold & Direct" or "Serious & Trusted" which have similar positioning without the same level of strictness.

---

Tips for Writing a Stronger Input

Good name generation lives or dies on your input. A few things that consistently help:

On your idea description:

Name the actual product category, not just a vague function

Include what makes it different or interesting

Mention the problem it solves, not just what it does

On your audience description:

Be specific about who the customer is — job title, life situation, behavior

If there are two audiences (e.g. a marketplace), name both

Avoid "everyone" — it tells the engine nothing

On your vibe:

If you're not sure, use "Surprise Me" — it performs well and maximizes yield

If you picked a strict vibe and got few results, try a looser one before assuming the engine is broken

The engine is only as good as the brief you give it. More specificity in, better names out.