Finding the right .eth name used to mean staring at a search box, typing guesses, and hoping something was available. You already know the name you want is taken. The interesting question is what you'd want instead — and that's exactly the problem keyword search is bad at solving.
AI Search on ens.tools flips the workflow. Instead of typing exact strings, you describe what you're looking for in plain English and let the search do the pattern-matching across a curated catalogue of more than 200,000 names.
Why plain-English search beats keyword guessing
Traditional ENS search is a lookup: you type vault.eth, it tells you it's registered, and you're back to square one. That's fine when you know the exact name. It falls apart when you're exploring — when you have a vibe, a theme, or a use case in mind but not a specific string.
AI Search is built for that exploratory mode. You can type things like:
- "short punchy names for a DeFi protocol"
- "three-letter names that sound like a wallet"
- "names related to gaming or esports"
- "palindromes under six characters"
The search interprets intent rather than matching literal characters, then surfaces candidates from the 200,000+ name catalogue that fit the description. You're browsing possibilities, not just confirming what's gone.
How to run a good AI Search
The quality of your results scales with the clarity of your prompt. A few practical habits help.
Start with the use case, not the word
Rather than searching for a single keyword, describe what the name is for. "A memorable name for a crypto newsletter" gives the search far more to work with than "crypto." Context lets it weigh tone, length, and theme together.
Layer in constraints
If you care about length, digits, or a particular pattern, say so. "Four-character names, letters only" or "names with a double letter" narrows the field fast. The catalogue already includes structured collections — 3-letter names, the 999 Club, palindromes — so pattern-based prompts land cleanly.
Iterate
Treat the first result set as a starting point. If everything skews too generic, add a constraint. If it's too narrow, loosen one. Because you're describing intent, small wording changes can meaningfully reshape what comes back. Two or three passes usually gets you to a shortlist worth saving.
From search result to shortlist
Once names start looking promising, you don't have to hold them in your head. Add candidates to your watchlist so you can compare them side by side and check back later. This matters because availability and status change: a name might be registered now but heading toward expiry, or already listed on the marketplace.
That's where AI Search connects to the rest of the platform. A name you like might be:
- Available to register directly on Ethereum mainnet.
- Listed for sale on the non-custodial marketplace, where you can see the asking terms in the live Activity feed.
- Approaching expiry, which you can track in the Expiry Explorer.
The point isn't to rush a decision. It's that AI Search feeds naturally into tools that show you the full state of each name, so your shortlist is grounded in real, current data rather than a stale snapshot.
Searching by theme and collection
One underused move: let AI Search pull from the platform's Collections. If you ask for "names in the 10k Club" or "three-letter dictionary words," the search can lean on those curated groupings. Collections are essentially pre-organised slices of the namespace — numeric clubs, dictionary words, palindromes — and describing them in your prompt gives the AI a clean structure to draw from.
This is handy when you care about a name's category as much as the name itself. Membership in a recognised collection is part of what makes a name legible to other people in the ecosystem, and searching by collection theme surfaces those names without you needing to memorise every club's rules.
Common mistakes that weaken your search
A few habits quietly limit what AI Search can do for you, and they're easy to avoid once you're aware of them.
Searching for a single word
Typing one keyword throws away most of the search's advantage. A single word carries almost no context, so the results tend to be broad and generic. The more you describe the purpose, tone, and constraints, the sharper the results.
Front-loading every constraint at once
Piling five requirements into the first prompt often over-narrows the field before you've seen what's possible. Start slightly broad, look at what comes back, then tighten. You'll usually discover a direction you hadn't considered — which is the entire reason to explore rather than look up.
Not saving anything
It's easy to run a great search, spot two names you like, then lose them by refining the prompt. Add promising names to your watchlist before you iterate, so each pass builds your shortlist instead of replacing it.
A quick workflow to try
Here's a repeatable loop that works well:
- Describe the goal — one sentence about what the name is for and any hard constraints.
- Scan the results — note two or three that feel right.
- Add them to your watchlist — so nothing gets lost.
- Check status — is each one available, listed, or expiring soon?
- Refine and repeat — adjust the prompt and run it again to widen your options.
Fifteen minutes of this usually beats an afternoon of typing guesses into a plain search box.
Where AI Search fits the bigger toolkit
AI Search is the front door, but it's most powerful as part of a loop. The catalogue it draws on is curated rather than a raw dump of the entire namespace, which is why the results tend to be usable rather than noisy. Once you've found candidates, the Expiry Explorer tells you which are approaching renewal, the Activity feed shows whether any are currently listed or attracting offers, and Collections let you jump sideways into the club or category a name belongs to. Search finds the name; the surrounding tools tell you its full story. Getting comfortable moving between them is what turns a one-off search into a genuinely fast way to explore ENS.
Try AI Search on ens.tools
The fastest way to understand AI Search is to use it. It's free to explore, it runs over a curated 200,000+ name catalogue, and it turns "I'll know it when I see it" into an actual search you can run. Head to ens.tools, open AI Search, and describe the .eth name you're picturing — then let the catalogue do the work.