How to conduct a clearance search on the IP India public portal

1. Introduction: The Cheapest Insurance in Brand Protection

A clearance search costs nothing on the IP India portal. Not doing one can cost a rebrand.

The arithmetic is brutal and entirely predictable. A founder spends eighteen months and a marketing budget building a name, files Form TM-A, waits a year for examination, receives an objection under Section 11 citing a mark that was sitting in the public database the entire time, spends another year fighting it, loses, and starts over — with signage, packaging, domain, app store listing, invoices, and customer recognition all built around a name that was never available. The government filing fee was ₹4,500 or ₹9,000 per class. The rebrand is a different order of magnitude.

The Trade Marks Registry maintains a free, public, comprehensive database of every trade mark application filed in India, and it requires no registration, no login, and no payment. Anyone can search it. Yet the majority of Indian applicants either skip the search entirely or run a single exact-match query in one class, see nothing, and conclude the name is clear. That is not a clearance search. That is a spot check that misses most of the ways a mark actually fails.

This guide sets out how to run a proper clearance search on the IP India public portal: what to do before you open the site, how each of the search modes works and what each one misses, how to construct the variant list that determines whether the search is worth anything, how to read result statuses correctly, what the portal cannot tell you, and how to convert the raw results into an actual risk assessment.

This is general information on Indian trade mark practice as of 2026, not legal advice. Portal interfaces and government fees change; verify current details on ipindia.gov.in. A clearance search is not a substitute for a professional availability opinion where the commercial stakes are significant.


2. What a Clearance Search Is Actually For

A clearance search answers two distinct questions, and confusing them is the most common conceptual error.

Question 1 — Registrability: will the Registry accept this mark? This is about objections under:

  • Section 9 (absolute grounds): is the mark descriptive, generic, laudatory, deceptive, or otherwise devoid of distinctive character?
  • Section 11(1) (relative grounds): is there an earlier identical or similar mark for identical or similar goods or services, creating a likelihood of confusion?
  • Section 11(2): is there an earlier well-known mark, protection of which extends even to dissimilar goods or services?

Question 2 — Freedom to use: can somebody sue me even if I get registered? Registration is not a licence to use. A registered proprietor can still be restrained by:

  • Section 34, which protects a continuous prior user against the registered proprietor
  • Section 27(2), which preserves the common law action for passing off in respect of unregistered marks
  • Section 11(3), under which a mark cannot be registered if its use in India is liable to be prevented by the law of passing off or copyright

The IP India portal answers Question 1 well. It answers Question 2 only partially, because unregistered prior users do not appear on the Register at all. A search confined to the portal is therefore always incomplete. Section 12 of the guide below deals with the supplementary searches that close that gap.


3. The Portal: What It Is and What It Contains

3.1 Access

The Trade Marks Registry’s public search is reached either directly at the Trade Marks Registry search address (tmrsearch.ipindia.gov.in/tmrpublicsearch/) or by navigating from ipindia.gov.in or ipindiaonline.gov.inTrade MarksPublic Search. Navigating from the main site typically produces an interstitial prompt confirming that you are being directed to an external service; confirm and proceed.

It is free, requires no account, and imposes no usage limits.

3.2 Coverage

The database contains every trade mark application filed in India — pending, objected, opposed, advertised, registered, refused, withdrawn, abandoned and removed — together with the mark, the class, the proprietor, the address, the date of application, the goods/services specification, the user date claimed, and the current status with its history.

International registrations designating India under the Madrid Protocol also feed into the Indian records and are searchable, though there can be a lag between WIPO records and Indian database reflection.

3.3 The Search Modes

The portal offers, at present:

Mode What it searches Primary use
Wordmark The text of the mark Brand names, product names, taglines
Phonetic Marks that sound similar, per the Registry’s algorithm Aural conflicts across different spellings
Vienna Code Figurative (device) elements of logos, by Vienna Classification code Logo and device conflicts
Proprietor / applicant name Marks held by a named person or entity Competitor portfolio mapping, chain-of-title checks

Additionally, on 18 September 2024, DPIIT and the Ministry of Commerce and Industry launched an AI and ML-based trade mark search technology, alongside the IP Saarthi generative AI chatbot for user assistance. The AI search surfaces conceptually and phonetically similar marks that a plain text query can miss, supports image upload for logo comparison, and can compare marks across scripts and languages — a meaningful capability in a country with as many writing systems as India. It should be treated as a valuable second pass, not as a replacement for a structured manual search.

3.4 Adjacent Databases on the Same Portal

Do not stop at the trade mark search. The IP India ecosystem also offers:

  • TM Classification / “Search Goods & Services” — enter a plain description of your product or service and the tool maps it to the correct Nice class. Use this before searching, not after.
  • Trade Marks Journal — searchable and downloadable; contains marks advertised for opposition.
  • List of well-known trade marks — determined under Sections 11(6)–11(9) read with Rule 124 of the Trade Marks Rules, 2017. Essential for cross-class risk.
  • Application status / e-Register — full prosecution history of any application number.
  • Patent (InPASS), Design and GI databases — relevant where the brand element overlaps with a design or a geographical indication.

4. Step 0: Preparation Before You Open the Portal

Most bad searches fail before the first query is typed. Do this first.

4.1 Fix the Mark Precisely

Decide exactly what you are searching:

  • Word mark — the name in plain text, no styling. Protects the name in any font or style.
  • Device mark — a logo, symbol, or artistic element.
  • Composite mark — name plus logo together.

If your brand has both a name and a logo, these are two marks and require two searches (and ultimately two applications). Search the word element and the figurative element separately.

4.2 Identify Your Classes

Use the TM Classification tool. Enter what you actually sell.

Two rules:

  • Search every class you will operate in, plus adjacent classes. A clothing brand searches Class 25, but also Class 35 (retail), and possibly Class 18 (bags) and Class 14 (jewellery/watches) if the range extends.
  • Search “allied and cognate” classes even if you will not file there. A mark registered in Class 30 (packaged foods) is a real risk to your Class 43 (restaurant services) plan, because courts and the Registry recognise conflict across related trades even where classes differ.

Note also that the 13th edition of the Nice Classification took effect on 1 January 2026, moving certain items between classes. Confirm the current classification of your goods rather than relying on an older list.

4.3 Build the Variant List

This is the single highest-value step, and the one most often skipped. Before searching, write down every string a conflicting mark might take. For a hypothetical mark KWIKFIX:

Spelling variants: QUICKFIX, QUIKFIX, KWICKFIX, KWIKFIKS, QWIKFIX Consonant substitutions: C ↔ K ↔ Q, S ↔ Z ↔ C, F ↔ PH, J ↔ G, V ↔ W, D ↔ T, B ↔ P Vowel substitutions and drops: KWEEKFIX, KWIKFEX, KWKFIX Doubled/single consonants: KWIKKFIX, KWIKFIXX Separations: KWIK FIX, KWIK-FIX, K-WIK FIX Prefixes and suffixes: SUPERKWIKFIX, KWIKFIXPRO, NEW KWIKFIX Roots alone: KWIK, KWIC, QUIK, FIX Transliterations: the Devanagari, Tamil, Telugu, Bengali, Gujarati or Urdu rendering of the mark, and the English rendering of an Indian-language mark Translations: the Hindi or regional-language equivalent of an English mark, and vice versa (relevant because the doctrine of foreign equivalents is applied in India)

A list of 15–30 variants is normal for a two-syllable mark. Searching only the exact mark is close to useless.

4.4 Draft a Provisional Specification

Write out, in plain terms, the goods or services. You will need this to assess whether a hit is in a genuinely conflicting trade or merely shares a class number.


5. The Wordmark Search: Step by Step

5.1 The Fields

On the public search page:

  1. Search Type: select Wordmark.
  2. Wordmark: enter the string.
  3. Class: enter one class number. The portal searches one class at a time. This is the portal’s most significant practical limitation and the reason searches take longer than people expect. Budget for (number of variants) × (number of classes) queries.
  4. Match logic / prefix: choose Start With, Contains, or Match With.
  5. Enter the captcha and run the search.

5.2 Understanding the Three Match Options

Option Returns Use it for
Match With Exact matches only The first, fastest check — is the identical mark already taken in this class?
Start With Marks beginning with your string Catching prefixed and suffixed marks: searching “KWIK” returns KWIKFIX, KWIKSERVE, KWIKCLEAN
Contains Marks containing your string anywhere The workhorse. Searching “KWIK” returns SUPERKWIK, MYKWIK, KWIKFIX — everything

5.3 The Recommended Sequence

For each class, in order:

  1. Match With on the full mark — the go/no-go check.
  2. Start With on the full mark — catches the mark plus suffixes.
  3. Contains on the full mark — catches the mark embedded in a longer mark.
  4. Contains on the distinctive root — this is the query that finds most real conflicts. For KWIKFIX, search “KWIK” and separately “FIX” (though FIX will be noisy).
  5. Contains on each of your spelling variants.

5.4 The Truncation Principle

Search short, not long. Long strings return almost nothing, which feels reassuring and is misleading. The distinctive part of the mark — usually the first syllable or two — is where conflicts cluster, because that is where consumer recollection concentrates and where the Registry’s citation practice focuses. If your mark is VEDANTIKA, search “VEDANT”, “VEDAN”, “VEDANTI” — not just “VEDANTIKA”.

5.5 Handling Descriptive and Common Elements

If your mark contains a common word (SUN, ROYAL, SHREE, GOLD, PRIME, TECH, CARE, MART), searching that element in a busy class will return hundreds of hits. Do not skip it. Instead:

  • Note the overall volume — a crowded field means the common element is weakly protected, which cuts both ways: your objection risk is lower, and your own exclusivity will be narrow.
  • Focus on hits where the whole mark is similar, not just the common element.
  • Check whether any hit is a well-known mark or a very large proprietor.

6. The Phonetic Search: Step by Step

6.1 Why It Matters Legally

Section 11(1)(b) turns on similarity and likelihood of confusion. Indian courts have consistently held that aural similarity is often decisive, particularly for goods bought across a counter, ordered verbally, or purchased by consumers who may not read English. Phonetic similarity has been treated as especially important in pharmaceuticals, where the consequences of confusion are severe.

A mark spelled entirely differently from yours can therefore block you. KOKA and COCA. FEVIKWIK and PHEVIQUICK. LAKME and LACMAY. The wordmark search will not find these. The phonetic search is designed to.

6.2 Running It

Select Phonetic as the search type, enter the mark, select the class, and run. The mechanics mirror the wordmark search, except that the prefix/match logic is not applied in the same way — the algorithm works on pronunciation-based similarity as the Registry’s system interprets it.

6.3 Its Limitations — and the Manual Workaround

The Registry’s phonetic algorithm is useful but not exhaustive. It is built on English-language phonetic logic and does not reliably capture Indian-language pronunciation patterns, regional accent variation, or transliteration equivalences.

Therefore, run the phonetic search and then run wordmark “Contains” searches on your manually constructed phonetic variants. The two together are far stronger than either alone. Specific patterns to construct manually in the Indian context:

  • V/W and B/V confusion (common across several Indian languages)
  • Aspirated/unaspirated pairs: T/TH, D/DH, P/PH, K/KH, B/BH, G/GH
  • Retroflex/dental: T/TT, D/DD
  • S/SH/SS
  • Terminal vowel presence or absence: RAMA / RAM, SITA / SEETA
  • Devanagari and regional-script transliterations of an English mark, and the reverse
  • Hindi/regional translations of an English word mark

6.4 Skipping Phonetic Search Is the Classic Beginner Error

For invented, stylised, or creatively spelled brand names — precisely the names startups favour — the phonetic dimension is where the risk lives. A creatively misspelled version of an existing mark is not a workaround; it is a near-guaranteed objection.


7. The Vienna Code Search: Step by Step (Device Marks)

7.1 What Vienna Classification Is

The Vienna Classification is an international system for indexing the figurative elements of marks — the shapes, animals, objects, human figures, geometric forms, and abstract designs in a logo. Codes are hierarchical, structured as Category → Division → Section, expressed numerically (for example, a code in the form 3.1.6 or similar, depending on the element).

Because it indexes what a logo depicts rather than what it says, it is the only way to find visually similar logos bearing entirely different names.

7.2 Running It

  1. Select Vienna Code as the search type.
  2. Enter the relevant Vienna code(s) for the figurative elements in your logo.
  3. Enter the class.
  4. Run and review the device marks returned.

7.3 Finding Your Codes

Break the logo into its visual components and identify a code for each: a lion, a stylised leaf, a shield outline, three concentric circles, a stylised letter within a square. WIPO publishes the Vienna Classification, and several free lookup tools reproduce it. Search each element’s code separately.

7.4 Its Real Limitations

Vienna coding in Indian records is assigned by the Registry, not the applicant, and coding practice is inconsistent. Two examiners can code the same device differently, and older records may be coded sparsely or not at all. You will also see the status “Send to Vienna Codification” on applications awaiting coding — those marks are effectively invisible to a Vienna search until coded.

This means a Vienna search is necessary but not sufficient for device clearance. Supplement it with:

  • The AI-based image search on the portal, uploading your logo
  • A reverse image search on general web search engines
  • A visual scan of the results returned by wordmark searches in your class

8. The Proprietor Search

Search by applicant or proprietor name to:

  • Map a competitor’s entire portfolio — every mark they hold, in every class, with statuses. This tells you their expansion plans before they announce them.
  • Check chain of title — whether the mark you are worried about has been assigned, and to whom.
  • Assess an opponent’s litigation posture — a proprietor with dozens of oppositions filed against third parties is an opponent you should plan for.

This is also the search to run before naming a company, before an acquisition, and before sending or responding to a cease-and-desist notice.


9. Reading the Results Correctly

9.1 The Result Columns

Typical fields returned: application number, mark (word and/or image), class, proprietor name, date of application, and status. Click through to the full record for the specification, user date, address, agent, and prosecution history.

9.2 The Status Glossary — and Why It Matters

Status What it means Risk to you
New Application / Formalities Chk Pass Filed, awaiting examination High — earlier filing date beats yours
Formalities Chk Fail Deficiency in filing Moderate — may be cured
Marked for Exam / Send to Vienna Codification In processing queue High
Objected Examination report issued High — objections are frequently overcome
Accepted & Advertised Published in the Journal Very high — heading to registration
Opposed Under opposition Moderate to high — outcome uncertain
Registered On the Register Highest
Refused Rejected by the Registrar Lower — but check for appeal, and note the mark may still be in use
Withdrawn / Abandoned Applicant did not pursue it Lower — but the mark may still be used in the market
Removed / Registration Expired Not renewed Lower — but restorable, and common law rights survive

9.3 Three Rules for Interpreting Status

Rule 1: Pending marks are as dangerous as registered ones. Under Section 23, registration relates back to the date of application. A pending application with an earlier filing date will be cited against you, and if it registers, it registers with priority over you. Never dismiss a hit because it says “Objected” or “New Application.”

Rule 2: Abandoned and refused marks still matter. Registration status has nothing to do with market use. A party whose application was abandoned may still be trading under the mark, may still have prior-user rights under Section 34, and may still sue you for passing off under Section 27(2). Look the proprietor up online.

Rule 3: Expired registrations can come back. Under Section 25, a registration lasts ten years and is renewable. Where renewal is missed, the Act and Rules provide for renewal with surcharge and for restoration within a prescribed window after expiry. A mark showing “Removed” may be restored and revived while your application is pending.

9.4 What to Extract from Every Serious Hit

For each hit that is genuinely similar, record:

  • Application number and current status
  • Exact mark as filed
  • Class and the actual specification of goods/services (not just the class number)
  • Proprietor name and address
  • Date of application
  • User date claimed
  • Whether the proprietor appears to be actively trading

The specification matters more than the class. Two marks in Class 9 — one for industrial safety helmets, one for mobile applications — are not in conflict despite sharing a class.


10. What the IP India Portal Cannot Tell You

A portal-only search is incomplete. Six categories of risk live outside the Register.

10.1 Unregistered Prior Users

India protects prior users powerfully. Section 34 preserves the rights of a continuous prior user against a registered proprietor, and the Supreme Court has repeatedly affirmed that a prior user’s rights are superior to those of a subsequent registrant. A business trading under your proposed name for fifteen years without ever filing an application will not appear anywhere on the portal — and can still stop you.

Search: general web search engines, Google Maps and local listings, e-commerce marketplaces (Amazon, Flipkart, Meesho), food delivery and aggregator platforms, app stores, YouTube, Instagram and Facebook handles, and industry trade directories.

10.2 Company and Business Names

Search: the MCA company/LLP name availability database; Udyam registration records; state Shop & Establishment records where accessible; GST taxpayer search. A registered company name is not a trade mark, but it is evidence of use and a source of dispute.

10.3 Domain Names and Digital Handles

Search: WHOIS records for .com, .in, .co.in; availability of the domain; social media handle availability across platforms. A blocked domain does not defeat registrability, but it materially affects whether the brand is commercially usable.

10.4 The Trade Marks Journal

Marks advertised in recent Journal issues are in the four-month opposition window and represent imminent registrations. Cross-check recent issues, particularly if your portal search returns “Accepted & Advertised” hits.

10.5 Well-Known Marks

The list of marks determined to be well-known under Sections 11(6)–11(9) and Rule 124 is published by the Registry. These marks are protected across classes under Section 11(2). If your proposed mark resembles one, class differentiation will not save you.

10.6 International and Related Registers

  • WIPO Global Brand Database — aggregates data across many jurisdictions; useful for foreign marks entering India, though not always synchronised with the Indian Register.
  • Madrid Protocol designations of India — searchable, with possible lag.
  • Copyright Register — relevant where the logo artwork may be someone else’s artistic work; Section 11(3) permits refusal where use would be prevented by copyright law.
  • GI Registry — critical for food, beverage, textile and handicraft brands using place names.

11. From Search Results to Risk Assessment

Finding hits is mechanical. Assessing them is judgement. Work through this framework.

11.1 The Four Scenarios

Scenario A — Identical mark, identical/similar goods, active status. Stop. Choose a different mark. No amount of stylisation, spelling variation, or class manoeuvring fixes this.

Scenario B — Similar mark, identical/similar goods. This is the difficult zone and the reason professional opinions exist. Assess:

  • Visual similarity — how the marks look side by side and, more importantly, in sequence
  • Aural similarity — how they sound when spoken by an ordinary consumer
  • Conceptual similarity — whether they convey the same idea
  • Doctrine of imperfect recollection — the comparison is not side-by-side; it is against a consumer’s hazy memory
  • Anti-dissection rule — marks are compared as wholes, though the dominant or essential feature carries greater weight
  • Nature of the goods and consumer — a casual FMCG purchase by an unsophisticated buyer demands a stricter standard than a high-value B2B capital equipment sale
  • Trade channel overlap — same shelf, same platform, same distributor?

Scenario C — Identical mark, different and unrelated goods. Often survivable, because trade mark rights are class- and trade-specific. But check whether the earlier mark is well-known (Section 11(2)) or whether the goods are “allied and cognate” — a concept Indian tribunals apply generously.

Scenario D — Crowded field. Many similar marks coexisting in the class indicates a weak, diluted element. Your objection risk falls, but so does the exclusivity you will obtain. A mark you can register but cannot enforce is a limited asset.

11.2 Also Screen Your Own Mark Under Section 9

Before celebrating a clean Section 11 result, check the absolute grounds. Is the mark:

  • Descriptive of the goods, their quality, quantity, purpose, value, or geographical origin?
  • Generic or customary in the trade?
  • Laudatory (BEST, SUPREME, PREMIUM)?
  • Deceptive or likely to cause confusion as to nature or origin?
  • A surname or common personal name without acquired distinctiveness?
  • Prohibited under the Emblems and Names (Prevention of Improper Use) Act, 1950?

A perfectly unique but wholly descriptive mark will fail on Section 9 no matter how clean the Section 11 search was.


12. Document the Search

Save your search record. Screenshot each query with the search string, class, date, and results visible; compile them into a dated PDF; store it with the brand file.

Why this matters practically:

  • It evidences bona fide adoption. Where an infringement defendant is shown to have adopted dishonestly, Indian courts respond with injunctions and, increasingly, punitive damages. A documented pre-adoption clearance search is the cleanest rebuttal to an allegation of dishonest adoption.
  • It supports an honest concurrent use claim under Section 12, if one ever becomes necessary.
  • It creates an internal audit trail showing that the decision to adopt was informed.
  • It gives your attorney a starting point rather than making them redo the work.

13. Practical Troubleshooting

Portal not loading. The Registry takes e-services offline for scheduled maintenance and posts notices on ipindia.gov.in. Check for a notice before assuming a fault.

Odd rendering or non-functioning buttons. The Registry portals were built for older browsers. Clear cache, disable aggressive ad-blockers, or try a different browser.

Captcha failures. Common; retype carefully and refresh.

No results at all for a common word. Verify that the class field is populated correctly and that the match logic is set as intended. A blank or wrong class silently narrows or breaks the query.

Results feel too clean. They usually are. If a two-syllable English-derived mark in Class 25 returns nothing on a “Contains” search of its root, re-check your inputs before believing it.


14. The Master Workflow Checklist

Preparation

  • Mark fixed and separated into word and device elements
  • Classes identified using the TM Classification tool (current Nice edition)
  • Adjacent and allied classes listed
  • Variant list built (spelling, phonetic, transliteration, translation, prefixes/suffixes, roots)
  • Provisional specification drafted

Portal searches — repeat for every class

  • Wordmark, Match With, full mark
  • Wordmark, Start With, full mark
  • Wordmark, Contains, full mark
  • Wordmark, Contains, distinctive root
  • Wordmark, Contains, each spelling variant
  • Phonetic search, full mark
  • Wordmark, Contains, each manually constructed phonetic variant
  • Vienna Code search for each figurative element (device marks)
  • AI/ML search as a second pass, including image upload for logos
  • Proprietor search on any concerning hit

Beyond the portal

  • Well-known marks list
  • Recent Trade Marks Journal issues
  • MCA company/LLP name search
  • Domain and social handle availability
  • Web, e-commerce, app store and local listings search for unregistered users
  • GI Registry (if place-name or produce-linked)
  • Copyright considerations for logo artwork
  • WIPO Global Brand Database (if international plans exist)

Assessment and record

  • Section 9 self-screen completed
  • Each serious hit assessed on visual, aural and conceptual similarity
  • Specifications (not just class numbers) compared
  • Search record compiled, dated and stored

15. When to Escalate to a Professional Search

Do it yourself where the mark is invented, the class is uncrowded, and the commercial exposure is modest.

Get a professional availability opinion where:

  • The brand will carry significant marketing investment
  • The mark is suggestive or contains descriptive elements (Section 9 risk requires judgement)
  • The relevant class is crowded — pharmaceuticals (Class 5), apparel (Class 25), IT services (Class 42), retail (Class 35), food (Classes 29, 30, 43)
  • Your search returned similar-but-not-identical marks and you cannot confidently assess them
  • Investors, lenders, or acquirers will conduct IP due diligence
  • International filing through the Madrid Protocol is planned
  • A competitor with a history of oppositions operates in your space

The portal gives you data. An opinion gives you a judgement on likelihood of confusion under Section 11, and that judgement is what actually determines whether you should adopt the mark.


16. Conclusion

The IP India public portal is one of the most useful free resources available to Indian businesses, and it is chronically underused. The database is complete, the access is unrestricted, and the search is instantaneous. What it demands is method.

The method reduces to a few disciplines. Identify the right classes before searching, not after. Build a variant list, because conflicts hide in spellings you have not thought of. Use Contains on the root, not Match With on the full mark. Always run the phonetic search, and then run the phonetic variants manually because the algorithm is not exhaustive. Search Vienna codes for logos, and back them up with the AI image search. Treat pending applications as dangerous and abandoned ones as still potentially in use. Then step outside the portal, because the prior user who never filed is the risk the Register cannot show you.

An hour spent on this before adopting a name is the highest-return hour in the entire brand-building process. The alternative — discovering the conflict eighteen months later in an examination report, or three years later in a cease-and-desist notice — costs the name itself.


17. Frequently Asked Questions (FAQs)

1. Where exactly is the IP India trade mark public search?

A. It sits at the Trade Marks Registry’s public search address, tmrsearch.ipindia.gov.in/tmrpublicsearch/, and can also be reached from ipindia.gov.in or ipindiaonline.gov.in by going to Trade Marks → Public Search. Navigating from the main site usually produces a prompt confirming redirection to an external service; confirm and continue.

2. Is the search free? Do I need to create an account?

A. It is completely free, requires no registration or login, and has no usage limit. Anyone can run unlimited searches.

3. What search types does the portal offer?

A. Wordmark (text of the mark), Phonetic (marks that sound similar), Vienna Code (figurative elements of logos), and proprietor/applicant name. Since September 2024 the Registry has also offered an AI and ML-based search that finds conceptually and phonetically similar marks and supports image upload, along with the IP Saarthi chatbot for guidance.

4. Can I search multiple classes at once?

A. No. The portal searches one class at a time; you must re-enter the class for each query. This is the portal’s most significant practical limitation. Plan for a number of queries equal to your variant count multiplied by your class count.

5. What is the difference between “Start With”, “Contains” and “Match With”?

A. “Match With” returns only exact matches. “Start With” returns marks beginning with your string, so searching KWIK returns KWIKFIX and KWIKSERVE. “Contains” returns marks with your string anywhere in them, so KWIK returns SUPERKWIK and MYKWIK as well. “Contains” is the most important of the three and the one that finds most real conflicts.

6. Should I search my full brand name or just part of it?

A. Search the distinctive root using “Contains”, not just the full mark. Long strings return almost nothing, which feels reassuring and is misleading. For VEDANTIKA, search VEDANT, VEDAN and VEDANTI as well as the full mark. Conflicts cluster at the beginning of marks, because that is where consumer recollection concentrates.

7. What is a phonetic search and why can’t I skip it?

A. It finds marks that sound like yours even when spelled completely differently — KOKA against COCA, or KWIK against QUICK. Section 11(1)(b) turns on likelihood of confusion, and Indian courts treat aural similarity as often decisive, especially for goods bought verbally or across a counter. Skipping it is the classic first-timer error, particularly for creatively spelled startup names.

8. Is the Registry’s phonetic algorithm reliable on its own?

A. Not entirely. It is built on English-language phonetic logic and does not reliably capture Indian-language pronunciation patterns, regional accent variation or transliteration equivalences. Run it, then run wordmark “Contains” searches on manually built variants covering V/W and B/V confusion, aspirated pairs like T/TH and D/DH, S/SH variations, terminal vowel presence, and script transliterations.

9. How do I search for a logo?

A. Use the Vienna Code search. Break the logo into its visual components — a lion, a shield outline, three circles, a stylised letter in a square — find the Vienna Classification code for each element, enter the code with the class, and review the device marks returned. Also upload the logo to the AI-based image search as a second pass.

10. Are Vienna Code searches complete?

A. No. Vienna codes in Indian records are assigned by the Registry rather than the applicant, and coding practice varies between examiners. Marks showing the status “Send to Vienna Codification” are not yet coded and will not appear at all. Supplement Vienna searching with the AI image search, a reverse image search on the open web, and a visual scan of results from your wordmark searches in the same class.

11. What is the AI-based search and should I rely on it?

A. The AI and ML-based trade mark search was launched by DPIIT and the Ministry of Commerce and Industry on 18 September 2024, together with the IP Saarthi generative AI chatbot. It surfaces conceptually and phonetically similar marks that plain text queries miss, supports image upload, and compares marks across scripts. Use it as a valuable second pass, not as a replacement for a structured manual search.

12. A similar mark shows status “Objected”. Can I ignore it?

A. No. Objections are frequently overcome. More importantly, under Section 23 registration relates back to the date of application, so an earlier-filed application that eventually registers will have priority over you. Pending applications should be treated as live risks regardless of their present status.

13. A conflicting mark shows “Abandoned” or “Refused”. Am I safe?

A. Not necessarily. Register status has nothing to do with market use. A party whose application was abandoned may still be trading under the mark, may hold prior-user rights under Section 34, and may sue for passing off under Section 27(2), which preserves that remedy for unregistered marks. Look the proprietor up online before concluding anything.

14. What about a mark showing “Removed” or expired?

A. Registrations last ten years under Section 25 and are renewable. Where renewal is missed, the Act and Rules provide for renewal with surcharge and for restoration within a prescribed window after expiry. A “Removed” mark may be restored and revived while your own application is pending, so treat it as a residual risk rather than a dead entry.

15. Does a clean search in my class mean my mark is available?

A. No, for three reasons. First, protection can extend across classes where the earlier mark is well-known, under Section 11(2). Second, Indian tribunals apply the “allied and cognate goods” concept generously, so a different class number does not guarantee a different trade. Third, unregistered prior users do not appear on the Register at all and can still stop you.

16. How do I find unregistered users who aren’t on the portal?

A. Search outside it: general web search, Google Maps and local business listings, e-commerce marketplaces, food delivery and aggregator platforms, app stores, YouTube, Instagram and Facebook handles, and trade directories. Also run the MCA company and LLP name check, GST taxpayer search, and a WHOIS lookup on the relevant domains.

17. How many classes should I actually search?

A. Every class you will operate in, plus adjacent classes, plus classes covering allied and cognate trades even where you do not intend to file. A clothing brand searches Class 25 and Class 35 at minimum, and often Classes 18 and 14. A packaged food brand should search Class 43 even if it has no restaurant plans, because a conflicting restaurant mark can still be raised against it.

18. How do I know which class my product falls in?

A. Use the TM Classification tool on the IP India portal, under Trade Marks. Enter a plain description of what you sell and it maps to the Nice class. Do this before you search, not after. Note also that the 13th edition of the Nice Classification took effect on 1 January 2026 and moved certain items between classes, so verify against the current edition.

19. How many variants should I search?

A. For a typical two-syllable mark, 15 to 30 variants is normal. Cover spelling alternatives, consonant substitutions (C/K/Q, S/Z, F/PH, V/W), vowel changes and drops, doubled consonants, spaced and hyphenated forms, prefixed and suffixed versions, the root alone, transliterations into and out of Indian scripts, and translations into Hindi or the relevant regional language.

20. Does the doctrine of foreign equivalents apply in India?

A. Indian practice does consider translations and transliterations, which is why a Hindi or regional-language rendering of an existing English mark is not a safe workaround, and vice versa. Search both directions.

21. My mark is unique but describes what I sell. Is a clean search enough?

A. No. A clean search only clears the relative grounds under Section 11. You must separately screen your own mark against the absolute grounds in Section 9 — descriptiveness, genericness, laudatory terms, deceptive matter, common surnames, and marks prohibited under the Emblems and Names (Prevention of Improper Use) Act, 1950. A wholly descriptive mark fails no matter how unique it is.

22. Many similar marks already coexist in my class. Is that good or bad?

A. Both. A crowded field indicates the common element is diluted, which lowers your objection risk. It also means whatever exclusivity you obtain will be narrow, so you may register the mark but struggle to enforce it. Focus on hits where the mark as a whole resembles yours rather than hits sharing only the common element.

23. Should I compare classes or specifications?

A. Specifications. Two marks in Class 9 — one covering industrial safety helmets, one covering mobile applications — are not in real conflict despite sharing a class number. Always open the full record and read the actual goods or services before treating a hit as serious.

24. What is the well-known marks list and why does it matter for my search?

A. It is the Registry’s published list of marks determined to be well-known under Sections 11(6) to 11(9), applied for under Rule 124 of the Trade Marks Rules, 2017. Well-known marks are protected across classes under Section 11(2), so if your proposed mark resembles one, filing in an unrelated class will not protect you.

25. Should I check the Trade Marks Journal separately?

A. Yes, particularly if your search returns hits with status “Accepted & Advertised”. Those marks are in the four-month opposition window and are heading to registration. The Journal is searchable and downloadable from the portal.

26. Do I need to search international databases?

A. Search the WIPO Global Brand Database if you have export or Madrid Protocol plans, or if you suspect a foreign brand may enter India. Note that it is not always synchronised with the Indian Register, so a mark filed in India may not yet appear there — and vice versa.

27. Should I save my search results?

A. Yes. Screenshot each query showing the search string, class, date and results; compile them into a dated PDF and store it with the brand file. This evidences bona fide adoption, which is the cleanest rebuttal to an allegation of dishonest adoption in later litigation, supports any honest concurrent use claim under Section 12, and saves your attorney duplicated work.

28. The portal isn’t loading. What should I do?

A. Check ipindia.gov.in for a scheduled maintenance notice, since e-services are taken offline periodically. If there is no notice, clear your browser cache, disable aggressive ad-blockers, or try a different browser — the Registry portals were built for older browsers and can behave unpredictably on the newest ones.

29. Can I rely on a free third-party search tool instead?

A. Free commercial tools query the same underlying database and can be convenient, but they are not authoritative and may lag or filter results. Treat them as supplements. The IP India portal is the source of record; if a third-party tool and the official portal disagree, the portal governs.

30. When should I stop searching myself and hire someone?

A. When the marketing investment is significant, the mark is suggestive or partly descriptive, the class is crowded (Classes 5, 25, 30, 35, 42, 43), your search returned similar-but-not-identical marks you cannot confidently assess, investors or acquirers will run IP due diligence, or a competitor with a history of filing oppositions operates in your space. The portal gives you data; an availability opinion gives you a judgement on likelihood of confusion, and that judgement is what actually decides whether to adopt the mark.

31. Does a clearance search guarantee my application will be accepted?

A. No. It substantially reduces the risk of a Section 11 citation and lets you avoid obviously doomed marks, but examination outcomes involve the Examiner’s discretion, marks filed after your search but before your filing will not appear, and Section 9 objections turn on judgement. A search improves the odds; it does not eliminate risk.

32. How long should a proper clearance search take?

A. For a single-word mark in two classes, budget one to two hours if you follow the full variant and phonetic protocol. For a composite mark across four or five classes with Vienna coding, expect half a day. Anyone who tells you it takes five minutes has run one exact-match query and called it a search.