How Nice Classification Works: A Developer's Guide to Trademark Categories

You're parsing trademark data from an API and every record has a nice_classes field. Class 9, Class 42, Class 35. These numbers aren't arbitrary — they're the backbone of how trademarks are organized worldwide. If your application searches, files, or monitors trademarks, you need to understand what Nice classification means and how to work with it.

Nice classification is the international system for categorizing goods and services in trademark registrations. Two identical brand names can coexist if they're registered in different classes — "DELTA" is both an airline (Class 39) and a faucet company (Class 11). Getting the class wrong in a search or filing workflow means your users get incorrect results.

What Is Nice Classification?

The Nice classification system originates from the Nice Agreement, signed in 1957 in Nice, France, and administered by the World Intellectual Property Organization (WIPO). It divides all commercial goods and services into 45 classes: classes 1 through 34 cover goods, and classes 35 through 45 cover services.

Nearly every major trademark office uses Nice classification — USPTO, EUIPO, WIPO (Madrid), CIPO, IPOS, and hundreds more. When someone files a trademark, they specify which Nice classes the mark covers. When someone searches for conflicts, class overlap is one of the strongest signals of risk.

WIPO updates the system periodically, publishing a new edition roughly every five years with annual version updates in between. Each update may reclassify specific goods or services, add new entries, or refine class descriptions.

For developers, the practical takeaway: class data is a required dimension in almost every trademark operation your application will perform.

Why Developers Need to Understand Trademark Classes

If you're building anything that touches trademarks — a search tool, a filing platform, a brand monitoring service, a marketplace compliance system — Nice classification classes show up in three places:

In search results. Every trademark record includes the classes it's registered or applied in. A search for "APEX" might return 200+ results, but filtering by Class 9 (software) narrows it to the ones that matter for a tech company.

In filing workflows. Trademark applications require applicants to select one or more Nice classes and list specific goods or services within each. Multi-class filings are common — a SaaS company might file in Class 9 (downloadable software), Class 35 (online retail services), and Class 42 (SaaS platform). Each additional class typically adds to the filing fee.

In risk assessment. Two identical marks in different classes generally coexist without issue. Two identical marks in the same class — or related classes — trigger a conflict. Your clearance or monitoring logic needs to account for class relationships, not just exact class matches.

All 45 Nice Classification Classes

The full list of trademark classes. Classes developers encounter most frequently are marked with an asterisk.

Goods (Classes 1-34)

Services (Classes 35-45)

For tech companies, Class 9 (downloadable software, electronic devices), Class 35 (online advertising platforms, data processing for business), and Class 42 (SaaS, cloud computing, software development services) come up most often. E-commerce platforms also work frequently with Class 25 (clothing) and Class 28 (toys/games).

Common Gotchas and Edge Cases

Software: Class 9, Class 42, or Both?

This distinction catches even experienced filers off guard:

• Class 9 covers downloadable software — applications installed on a device. Mobile apps, desktop software, firmware.
• Class 42 covers software as a service — cloud platforms, hosted applications, software development tools delivered as a service.

Most modern software companies file in both. If your product is a mobile app with a cloud backend, the app falls under Class 9 and the service under Class 42. Your filing workflow should guide users toward selecting both when applicable.

Class Headings vs. Specific Goods/Services

Each class has a heading (like "Scientific and technological services" for Class 42), but the heading is a summary, not a catch-all. Applicants must list specific goods or services within the class.

WIPO maintains an alphabetical list of goods and services with pre-approved terminology. Using terms from this list speeds up examination. Custom descriptions can trigger an office action asking for clarification.

Same Word, Different Trademark Classes

"Apple" in Class 31 (fresh fruits) is a completely different trademark context than "Apple" in Class 9 (computers). Your search interface needs to present class context prominently — without it, users draw wrong conclusions about conflicts.

Edition Changes

When WIPO publishes a new edition, some goods or services move between classes. Historical trademark data reflects the Nice classification at the time of filing. Your application should handle both current classification lookups and historical class assignments in older records.

Working with the Nice Classification API

Signa's Classifications endpoint lets you browse and search Nice classification data programmatically. List all classes:

curl https://api.signa.so/v1/classifications/nice \
  -H "Authorization: Bearer sig_live_..."

const classes = await signa.classifications.nice.list();

for (const cls of classes.data) {
  console.log(`Class ${cls.number}: ${cls.heading}`);
}

Look up what a specific class covers:

curl https://api.signa.so/v1/classifications/nice/9 \
  -H "Authorization: Bearer sig_live_..."

const class9 = await signa.classifications.nice.retrieve(9);

console.log(class9.heading);
// "Scientific and technological instruments, software, ..."

console.log(class9.includes);
// ["Downloadable software", "Computer hardware", ...]

Filter a trademark search by Nice class — where this becomes practical:

curl https://api.signa.so/v1/trademarks/search \
  -H "Authorization: Bearer sig_live_..." \
  -d '{
    "mark": { "text": "CloudForge", "strategies": ["exact", "phonetic"] },
    "filters": { "nice_classes": [9, 42] }
  }'

const results = await signa.trademarks.search({
  mark: { text: 'CloudForge', strategies: ['exact', 'phonetic'] },
  filters: { nice_classes: [9, 42] }
});

console.log(`Found ${results.total_count} results in classes 9 and 42`);

for (const tm of results.data) {
  console.log(`${tm.mark_text} — Classes: ${tm.nice_classes.join(', ')} — ${tm.status}`);
}

The nice_classes filter is critical for relevant results. A search for a common word without class filtering returns thousands of marks. Narrowing to relevant classes gives your users actionable signal.

Practical Example: Clearing a Software Trademark

You're building a clearance feature and a user wants to check "CloudForge" for a SaaS code deployment platform. Here's the thought process your application should encode:

Step 1: Identify the relevant Nice classification classes.

A SaaS deployment platform touches:

• Class 9 — downloadable CLI tools, desktop apps, SDKs
• Class 35 — only if the platform includes business management or analytics features
• Class 42 — the core SaaS platform itself

Step 2: Search across those classes.

Run the trademark search with nice_classes set to [9, 42] (and [35] if applicable). Use multiple search strategies — exact match finds "CLOUDFORGE" while phonetic matching catches similar-sounding marks like "KLOUDFORGE."

Step 3: Evaluate the results.

For each result in the relevant classes, consider:

• Is the mark live (registered or pending) or dead (abandoned, expired)?
• How similar is it — exact match, phonetic match, or fuzzy match?
• Is it in the same class or a related class?
• What jurisdiction — same one your user is filing in?

Step 4: Surface the risk.

Present results grouped by class, with similarity scores and status indicators. Marks that are live, in the same class, and phonetically similar represent the highest risk.

Coming in Phase 3, Signa will offer automated clearance screening that handles this entire workflow — multi-class, multi-jurisdiction risk scoring with a single API call. Today, the Search and Classifications endpoints provide the building blocks.

Consult a trademark attorney for legal guidance specific to your situation. Automated search results inform the analysis, but a qualified attorney should evaluate the legal risk before filing.

Explore Nice Classification with the Signa API

Try the Signa Classifications API to explore all 45 Nice classes programmatically. Browse classes, search for specific goods and services, and filter trademark searches by class to build classification-aware features into your product. Get your free API key at signa.so.