Classifications & Goods/Services

Why This Matters

When filing a trademark, you must specify exactly what goods or services your mark covers. This isn’t freeform: each trademark office maintains a list of pre-approved descriptions they accept. Using the right wording reduces the risk of examiner objections, which cause delays and extra costs. The challenge: different offices accept different wording. A description accepted at the USPTO may be rejected at the EUIPO, and vice versa. Signa solves this by aggregating 96,000+ pre-approved descriptions from the world’s two largest term databases, merged into a single searchable API.

The Nice Classification System

All trademark offices worldwide use the Nice Classification, a system of 45 classes that categorize goods and services:

Classes 1–34: Goods (physical products, chemicals, software, etc.)
Classes 35–45: Services (advertising, legal, education, etc.)

Use GET /v1/classifications?type=goods or ?type=services to filter, or fetch all 45 with no filter.

The 45 class headings are universal: every office uses them. But the specific descriptions within each class are where offices diverge. For example, Class 9 covers “scientific and electronic apparatus.” But when you file, you can’t just write “Class 9.” You need a specific description like:

“Downloadable computer software for managing cryptocurrency transactions”

Whether that exact wording is accepted depends on which office you’re filing at.

How Office Acceptance Works

There is no single universal list of accepted descriptions. Different offices accept different wording, with partial overlap between them. Signa aggregates terms from the Harmonised Database (maintained by EUIPO via TMClass and accepted by the EU national offices and a number of partner offices outside the EU) and the USPTO ID Manual (the US-specific database), and merges them at query time so you query one API instead of several office-specific tools. A term tagged is_harmonised: true comes from a harmonised list: either the EUIPO Harmonised Database (every TMClass term carries the flag) or an entry the USPTO ID Manual marks as TM5-harmonised (accepted at USPTO, EUIPO, JPO, KIPO and CNIPA). The flag does not by itself mean that every TM5 office, or the office you are filing at, accepts the wording: check accepted_offices for your filing office. Selecting an office-accepted term significantly reduces (though does not eliminate) the risk of an examiner objection, since office practice can still evolve.

Searching for Terms

Use the goods & services endpoint to search within a specific Nice class:

curl -G "https://api.signa.so/v1/goods-services" \
  -H "Authorization: Bearer sig_YOUR_KEY" \
  --data-urlencode "q=blockchain" \
  --data-urlencode "class=9"

const terms = await signa.goodsServices.list({
  q: "blockchain",
  class: 9,
});

for (const term of terms.data) {
  console.log(term.term, "→", term.accepted_offices);
}

Example Response

{
  "object": "list",
  "data": [
    {
      "term": "Downloadable computer software for managing cryptocurrency transactions on a blockchain",
      "term_key": "downloadable computer software for managing cryptocurrency transactions on a blockchain",
      "is_harmonised": false,
      "accepted_offices": ["USPTO"],
      "source": "uspto_idm",
      "language": "en"
    },
    {
      "term": "Downloadable software for blockchain-based inventory management",
      "term_key": "downloadable software for blockchain-based inventory management",
      "is_harmonised": false,
      "accepted_offices": ["USPTO"],
      "source": "uspto_idm",
      "language": "en"
    }
  ],
  "has_more": false,
  "pagination": { "cursor": null }
}

Understanding the Response Fields

Field	What it tells you
`term`	The exact description text. Use this wording in your filing.
`accepted_offices`	Which offices accept this exact wording. If you see `['USPTO', 'EUIPO']`, it works at both.
`is_harmonised`	`true` = the term comes from a harmonised list (the EUIPO Harmonised Database, or a USPTO ID Manual entry marked TM5-harmonised). It is not a per-office acceptance statement: read `accepted_offices` for that.
`source`	Where this term came from: `tmclass` (EUIPO), `uspto_idm` (USPTO).

Multi-Source Merging

When the same term exists in both the EUIPO and USPTO sources, Signa merges them into a single result with a combined accepted_offices list. For example, a term like “Acetone” that’s separately accepted at EUIPO and at USPTO (plus its TM5 partners) is returned as one row with accepted_offices: ['EUIPO', 'USPTO', 'JPO', 'KIPO', 'CNIPA']. This happens automatically: you always see one result per unique term with the full set of offices.

Common Patterns

Find descriptions for a product

Search with product keywords to find pre-approved descriptions:

# "What descriptions can I use for my AI legal research tool?"
curl -G "https://api.signa.so/v1/goods-services" \
  -H "Authorization: Bearer sig_YOUR_KEY" \
  --data-urlencode "q=legal research software" \
  --data-urlencode "class=42"

Find terms accepted at a specific office

Filter the results client-side by accepted_offices:

const terms = await signa.goodsServices.list({
  q: "software",
  class: 9,
  limit: 100,
});

// Terms accepted at USPTO
const usptoTerms = terms.data.filter(t =>
  t.accepted_offices.includes("USPTO")
);

// Terms accepted at BOTH USPTO and EUIPO
const bothOffices = terms.data.filter(t =>
  t.accepted_offices.includes("USPTO") &&
  t.accepted_offices.includes("EUIPO")
);

Find the safest wording for multi-jurisdiction filings

Use harmonised_only=true to get only terms with is_harmonised: true, then keep the ones whose accepted_offices lists every office you plan to file at:

curl -G "https://api.signa.so/v1/goods-services" \
  -H "Authorization: Bearer sig_YOUR_KEY" \
  --data-urlencode "q=computer software" \
  --data-urlencode "class=9" \
  --data-urlencode "harmonised_only=true"

A term accepted at every office in your filing plan is the lowest-risk wording, though office practice can still change.

Not sure which class?

Three options, in order of how much context you have:

Natural-language description: use Suggest Classifications. Pass a business description like "SaaS tool for HR teams" and it returns ranked classes with confidence scores and paste-ready terms.
Literal term lookup: omit class on the terms endpoint to see every class where that wording is accepted:
curl -G "https://api.signa.so/v1/goods-services" \ -H "Authorization: Bearer sig_YOUR_KEY" \ --data-urlencode "q=poster"
Inspect class_number on each result to discover relevant classes.
Class heading keyword: search the 45 class headings for an obvious keyword:
curl -G "https://api.signa.so/v1/classifications" \ -H "Authorization: Bearer sig_YOUR_KEY" \ --data-urlencode "q=software"
Software-related goods are typically in Class 9, while software-related services are in Class 42.

Goods and services on a record

Each item of a trademark’s classifications pairs a Nice class with the goods or services text the office published. Most offices state the text class by class. Some do not, so each item says how its text belongs to its class:

`goods_services_binding`	Meaning
`class`	The text is this class’s own: the office states it for this class, or the record has no other class of its kind.
`record`	The office states this text for the record’s goods (or services) as a whole, without naming a class, and the record has several goods classes (1 to 34) or several services classes (35 to 45). Signa repeats the text on each of them, so part of it may describe another of those classes.
`null`	The item has no text, or the binding is unknown. Signa sets `class` or `record` only from the office’s own list of goods and services statements, which it keeps so far for reprocessed Canadian (CIPO) records; on every other record the field is `null`.

Canada (CIPO)

CIPO numbers a record’s statements, such as goods (1), (2) and services (1), and lists its Nice classes separately. On many older filings, classified before Canada adopted the Nice Classification in 2019, the classes were added afterwards and no statement names its class. CIPO describes that classification data as “provided for information and searching purposes only”. Signa puts each such statement on every class of its kind, and marks the items record when there are several such classes. A statement that does name its class is kept on that class only and marked class, whatever the filing date. Signa is reprocessing Canadian records to this rule. Until a record is reprocessed, its items show null and a statement can sit on one class only or be missing. Every statement is searchable with goods_services_text, including a statement whose kind has no class on the record (a services statement on a record with only goods classes): it is in the searchable text but on no item of classifications. ?include=office_extensions returns the statements as CIPO numbers them, under goodsServicesStatements.

Data Coverage

Signa’s classification term database is updated automatically:

Source	Update frequency	Coverage
Nice Classification headings	Monthly	45 classes, ~10,000 base terms
EUIPO TMClass (Harmonised Database)	Quarterly	43,000+ terms accepted by the participating offices
USPTO ID Manual	Monthly	70,000+ terms accepted at USPTO
Vienna Classification (design codes)	Monthly	2,100+ visual element codes

Data freshness: The Nice taxonomy updates once per year (January). Individual terms are added throughout the year by each office. Signa syncs monthly to capture additions while keeping API response times fast.

Suggest Classifications

Turn a business description into ranked classes + filing-ready terms

List Goods & Services

Full endpoint reference with parameters and response schema

List Classifications

Get all 45 Nice class headings

Trademark Clearance

Use classifications in a full clearance workflow

Getting started

Core concepts

Build with AI

Data

Monitoring

Working with the API

Use cases

Classifications & Goods/Services

Why This Matters

The Nice Classification System

How Office Acceptance Works

Searching for Terms

Understanding the Response Fields

Multi-Source Merging

Common Patterns

Find descriptions for a product

Find terms accepted at a specific office

Find the safest wording for multi-jurisdiction filings

Not sure which class?

Goods and services on a record

Canada (CIPO)

Data Coverage

Suggest Classifications

List Goods & Services

List Classifications

Trademark Clearance

​Why This Matters

​The Nice Classification System

​How Office Acceptance Works

​Searching for Terms

​Understanding the Response Fields

​Multi-Source Merging

​Common Patterns

​Find descriptions for a product

​Find terms accepted at a specific office

​Find the safest wording for multi-jurisdiction filings

​Not sure which class?

​Goods and services on a record

​Canada (CIPO)

​Data Coverage

​Related

Suggest Classifications

List Goods & Services

List Classifications

Trademark Clearance

Why This Matters

The Nice Classification System

How Office Acceptance Works

Searching for Terms

Understanding the Response Fields

Multi-Source Merging

Common Patterns

Find descriptions for a product

Find terms accepted at a specific office

Find the safest wording for multi-jurisdiction filings

Not sure which class?

Goods and services on a record

Canada (CIPO)

Data Coverage

Related