AI-era due diligence for the names you bet on.
A name used to be worth how it sounds. Now, how AI reads it.
What the desk hands back.
Each is a real, full read you can click into and poke around — what AI already knows, the AI-era signals, the intent-routed verdict, the pricing band. The same read your own name would get.
Saturation-level recall. Every model knows it cold — see what the report does when the name is the category.
Clean brand, high attractor rate. The shape of a name the models reach for first when the topic comes up.
A dictionary word on an alt-TLD. Watch how the models confuse the brand with the concept.
An adjective on a product-TLD. See the intent-routed verdicts side by side.
Same name. Same price. Opposite call.
A comp tool hands every buyer the same number. But the recognition that makes a name a trap for the founder building on it is the exact thing an investor pays up for. Change who’s asking — the verdict flips.
A name the models already recognize is one a founder can’t manufacture at any price. That recognition is resale liquidity — the desk marks it up.
Same name. Same comp band. The number can’t tell these two buyers apart — the desk routes the verdict through which one you are.
Not a vibe. A finding.
A desk publishes its research. We don’t just read your one name — we study how the frontier models read names at scale. It’s why the read matters, and why no comp tool has it.
Made-up startup names, four open models, one question: what industry is this? The modal answer was “tech” — for nonsense, for real words, whatever you’re actually building.
read →599 names probed. Most get a confident, unanimous read; about 8% make the models disagree across industries — the highest-leverage band you can pick from.
read →The good ones left aren’t the names you’ll brainstorm — they’re the ones you have to invent. Here’s how the desk finds them.
read →“Can’t I just ask the model myself?”
You can ask one model, in plain words, and squint at what comes back. That’s a guess, not a read. Namedesk is a proprietary research instrument — it weighs many frontier models, computes AI-era signals you can’t pull from a chat box, and runs its own pricing engine to land on one number you can stand behind. The method is ours; the verdict is yours.
From 30¢ a read.
Pay as you go in credit packs that fit how much you research — one name, or a whole pipeline. The bigger the pack, the cheaper the read: price a hundred names for $30, less than a single overpay.
- On the AI shortlist — or invisiblenamedesk-onlyWhen a buyer asks AI for a tool in your space, does your name come up on its own? Miss it and you're invisible at the moment of choice. Almost no name clears this bar — we tell you if yours does.
- Catches what AI inventsnamedesk-onlyWhen the models confidently make up a backstory your name never had — a company that never existed, a history that isn't yours — we flag it before it misleads your customers.
- Prompt collision mapnamedesk-onlyWhich queries drag your name into a different brand's answer? Which neighbours are you competing with in the model's memory?
- Recall across eight modelsnamedesk-onlyHow much each of eight models — four open baselines plus the four frontier models people actually talk to (Claude, GPT-5, Gemini, Kimi) — already knows your name, on every read. One model's answer is an anecdote; eight is a signal.
- Intent-routed verdictnamedesk-onlyThe same read, written for who you are — build, monetize, invest, or explore. The verdict is framed for the call you're actually making.
- A proprietary pricing enginenamedesk-onlyOur own model factors the AI-era read into a defensible price band — not a blind average of stale registrar prices. Recomputed on every paid read.
- Brand signals, done properlyNine brand fundamentals — recall, confusion, brand, voice, pronounce, typeable, memorability, fit, TM risk — graded, not eyeballed.
Built for the person about to commit.
You just saw the flip. Whether you’re building on the name or buying to hold it, the desk reads it for the decision you’re actually making.
- Founders, about to commit to a name
A brand for the next decade is a one-way door. Run it first: which names the models will confuse for someone else, which ones they've already pinned to another company, and which are still clear of that memory.
- Investors pricing a portfolio
The AI-era read on every name you hold or chase — what the models already recognize, what they route around — across eight models, four open baselines plus the four frontier ones people actually talk to. Price up the proven names; stop renewing the dead weight. A desk for the whole pipeline, not one lookup at a time.
- Bulk flippers and registrar farms
No batch pipes, no API resellers, no pay-per-thousand lookups. This is a research desk, not a hose.
- SEO-first namers
If you're buying an exact-match keyword domain for search juice, there are cheaper tools. Our signal is brand fit in the LLM era, not SERP position.
What do the models already think of your name?
Sign up, get 10 free credits — enough for one research read or one coining shortlist. From 30¢ a read after that.