How to Check If a Brand Name Is Trademarked (2026 Guide — Free + Paid Options)

Why this matters in 2026

We've measured the trademark-check behavior of the first 1,200 founders who used Etymolt. The numbers are stark: 60% of independent founders never run a federal register check before launching a brand. Of the 40% who do run one, most stop at USPTO TESS Class 35 — “services” — and miss the relevant Nice class for software (9), consulting (42), education (41), or apparel (25).

The cost of skipping this check is non-linear. If you catch a conflict in week 1, you pivot for free. If you catch it in week 16 — like I did — you pay around $40,000 in legal fees, lose 11 weeks of go-to-market momentum, and redo the brand identity. Once is bad luck. Twice is a system problem.

The four levels of trademark check

Every trademark search you can run sits at one of four levels of rigor. Each level catches a different class of failure. The right level for your candidate depends on stakes, time, and budget.

Level 1 — USPTO TESS (free, ~30 minutes)

The Trademark Electronic Search System lives at uspto.gov/trademarks/search and is the canonical federal register for the United States. It is also a relic of the 1990s — searchable, but searchable in the way a microfilm archive is searchable. You must know what you're looking for. Here is the actual sequence.

1. Open TESS and choose Basic Word Mark Search (New User). Run your candidate. Note: TESS is case-insensitive but otherwise literal — “Stratagem” does not return “Strategem.”
2. Filter by Live records. Dead marks are abandoned, cancelled, or expired — they don't bar you from registering, though a recently-dead famous mark can still create a reputational risk.
3. Open every hit and scan the Goods and Services field. This is where the Nice class lives. Class 9 (software, downloadable apps, recorded media) and Class 42 (SaaS, software-as-a-service, technological consultancy) are the two classes that govern almost every AI startup. If the senior mark is in your class, you have a §2(d) collision risk.
4. Re-run the search using Free Form mode with a phonetic-equivalents query — TESS supports the [BI] truncation operator. For “Stratagem,” that means searching “strat$,” which catches Strata, Stratos, Stratagy, etc. Most §2(d) refusals cite phonetic-similar marks, not exact matches.

When you're done, you'll know whether a head-on collision exists. You will not know whether a senior-mark holder is litigious, whether a famous-mark dilution risk applies (Coca-Cola, Disney, Apple, Nike — all bar you regardless of your class), or whether your candidate is descriptive in the technical sense that triggers §2(e)(1). Those gaps are why level 1 is necessary but not sufficient.

Level 2 — Freemium tools ($0–$29, ~5 minutes)

Trademarkia, Markify, Trademark.com, and a half-dozen other operators wrap the USPTO public API behind a friendlier UI. They're fast and catch obvious collisions. The limits are well-known to attorneys and badly-known to founders.

• Phonetic search is shallow. Most freemium tools match on substring, not on actual phonetic distance. They miss the “Stratagy / Strategy” pair that any literate USPTO examiner would flag.
• Class filtering is opt-out, not opt-in. If you don't set the Nice class, you get a global search and a wall of irrelevant hits. Most founders tune out and miss the one hit that matters.
• No TTAB coverage. The Trademark Trial and Appeal Board is where opposition proceedings happen — the senior-mark holder who's about to oppose your application is visible in TTAB filings months before they file the opposition. Freemium tools rarely surface TTAB.

Freemium is a useful sanity-check after Level 1. It is not a replacement for either Level 1 (because it's shallower) or Level 3 (because it's slower and less cited).

Level 3 — API / automated check ($0 first 5, ~3 seconds)

The automated tier is what we built Etymolt to occupy. A single API call (POST /v1/verify) runs USPTO TESS, UKIPO, EUIPO, the Madrid Protocol register, TTAB, and a phonetic-distance scorer in parallel. The verdict comes back in under three seconds with every flag traced to a record number. Free for the first five clearances. The methodology is public.

The reason this tier matters is not speed (though sub-3s is genuinely useful when you're naming-storming on a Friday night). The reason this tier matters is coverage: a single API call hits every register that bars your registration, in parallel, with the same Nice-class logic an attorney would apply. Five free, no signup, every verdict permalinked.

Level 4 — Attorney consult ($250–$500/hr, ~1–4 hours)

The attorney tier is the only level that produces a defensible legal opinion. You want it when the stakes are high enough to justify the cost: a brand-defining product, a Series A pitch deck about to go out, a Class 9 or Class 42 software dispute, anything in Class 33 (alcoholic beverages) or Class 30 (food) where the famous-mark landscape is dense.

A clearance-search attorney will typically charge $1,500–$3,500 for a full opinion, which includes a literal opinion letter you can show to investors. The opinion letter is the unique deliverable; no Level 1–3 tool produces one. If a registration is later challenged, the opinion letter is what shifts the burden of proof in your favor.

The USPTO TESS walk-through (in detail)

If you only run one of these checks, run Level 1. Here is how to do it the way a clearance-search attorney would, in roughly thirty minutes, for free.

Open tmsearch.uspto.gov and choose the new search interface. The legacy interface still works but is being deprecated; the new one supports better filtering. Run your candidate as a single word, lowercase, with no quotation marks. You will get a result list with serial numbers, owners, and Nice classes.

Filter by Status: Live. There is a Live/Dead toggle at the top of every result list. Live includes registered marks and pending applications. Pending is important — it means the holder has filed but not yet been registered, and your application would be examined in the queue behind theirs.

For each Live hit, open the TSDR (Trademark Status and Document Retrieval) record. The Goods and Services field lists the Nice class. The Nice class is the single most important number on the record. Two marks can coexist if they're in different Nice classes and the holders' markets don't overlap — “Delta” is registered to Delta Air Lines (Class 39, transport) and to Delta Faucet (Class 11, plumbing) and the world keeps spinning.

The five Nice classes most relevant to modern software companies:

• Class 9 — downloadable software, recorded media, mobile apps. The killer class for any company whose product ships as a downloadable.
• Class 35 — advertising, business consulting, retail services. The default that founders search; rarely the one that matters.
• Class 41 — education, training, entertainment. Relevant for edtech, courses, content businesses.
• Class 42 — SaaS, software-as-a-service, scientific and technological services. The killer class for any company whose product is cloud-hosted.
• Class 45 — legal services, security services. Relevant for legal-tech and security companies.

When you find a Live mark in your candidate's Nice class, the next question is phonetic distance. The USPTO examiner applies the DuPont factors (a 13-point test for likelihood of confusion). The two factors that drive most refusals are similarity of the marks themselves (sight, sound, meaning) and similarity of the goods or services. If your candidate sounds like the senior mark and operates in the same Nice class, the refusal probability is high.

The TTAB lookup (the step everyone skips)

The Trademark Trial and Appeal Board hears opposition and cancellation proceedings. When a senior-mark holder opposes a junior application, the proceeding lives at ttabvue.uspto.gov. There are roughly 647,000 TTAB proceedings in the public record.

Why this matters: a senior-mark holder who's opposed three other applications in your Nice class is telegraphing that they will oppose yours. Their litigation pattern is a free signal that most founders never read.

The lookup: search TTABVUE for the senior mark's owner name. Scan the proceedings. If the owner has opposed multiple junior applications successfully, your refusal probability is materially higher than a casual TESS read would suggest. An attorney will check this as a matter of course; a Level 3 API check will surface it automatically; a Level 1 TESS read alone will not.

Madrid Protocol and international coverage

The Madrid Protocol is the WIPO-administered system for international trademark registration. If your brand is going to operate outside the US — which is true of roughly every internet-facing brand in 2026 — the Madrid register is the canonical source for prior international rights.

The pragmatic version: WIPO maintains a free search interface at wipo.int/madrid/monitor. Run your candidate. Filter by class. Note any registrations designating the markets you plan to enter — EU, UK, Canada, Australia, Japan, China are the usual suspects. An international mark designating the US gets imported into USPTO TESS automatically, but the reverse is not true.

The five questions to ask of every hit

When you find a senior mark that looks like a problem, the next step is triage. Five questions narrow the field from “every hit looks scary” to “this hit is actually a problem.”

1. Is it Live? Dead marks (abandoned, cancelled, expired) don't bar your registration. Recently-dead famous marks can still create reputational risk, but the legal bar is gone.
2. Same Nice class? Different Nice classes can coexist if the markets don't overlap. Software (Class 9 + 42) and physical hardware (Class 9 narrowly, Class 11/12 broadly) sometimes coexist; software and consulting (Class 35) usually don't.
3. Phonetic distance? The DuPont factors weigh sight, sound, and meaning. “Stratagem” vs. “Strategy” — different sight, similar sound, similar meaning — fails the test. “Apple” (the fruit) and “Apple” (the computer company) coexisted for decades because the markets were different.
4. Famous? Famous marks are protected against dilution regardless of class — the Trademark Dilution Revision Act of 2006. Coca-Cola, Disney, Apple, Nike, McDonald's, Google, Amazon, Microsoft. If your candidate sounds like any of them, the answer is no.
5. Litigious owner? TTABVUE will tell you. An owner who opposes three junior applications a year will probably oppose yours. An owner who has never opposed anyone will probably not.

When to consult an attorney

The four-level ladder above will catch most disasters, but there are five specific situations that justify a Level 4 attorney consult on top of whatever Level 1–3 work you've done.

• Class 9 or 42 software dispute. These are the two classes where phonetic-similar collisions are most common and most fatal. A clearance opinion is cheap insurance if you're about to spend material money on the brand.
• Class 33 alcoholic beverages. Dense famous-mark landscape, complex TTB labeling overlay. Not a DIY space.
• Class 30 food. Densely registered; descriptiveness traps everywhere (think “cold brew,” “artisan,” “handcrafted”).
• Class 25 apparel. Long-running litigation patterns; international rights matter more than in most classes.
• Series A coming up. Investor diligence will check trademark clearance. A clean opinion letter shortens the diligence cycle and protects the round.

For everything else, Levels 1–3 in sequence will catch ~80% of disasters and cost you between $0 and $29.

How Etymolt automates this

Etymolt is the Level 3 tier in the ladder above. A single API call to POST /v1/verify runs USPTO TESS, UKIPO, EUIPO, the Madrid register, and TTAB in parallel — plus four other axes (domain, cultural, sound symbolism, pronunciation resilience) that we'll cover in the companion guide The Five Ways a Brand Name Dies. The verdict comes back in under three seconds. Five free per IP. No signup. Every flag traces to a record number.

Clearance signal, not legal advice. Etymolt operates under the Bureau Model — we surface evidence; we're not a law firm. The verdict permalink stays with you forever and is citation-grade for investor diligence, partner intake, and your eventual Level 4 attorney consult.

Further reading

• The Five Ways a Brand Name Dies — trademark is one death, not all of them. The other four kill names at higher rates than founders expect.
• Why We Built Etymolt — A Founder's USPTO Office Action Story — the week-16 envelope, in detail. The $40K bill and the reset.
• The full five-axis methodology — what Etymolt checks, how we score, the registries we cite, the calibration we publish.