DOI Trademark Database Limits in Nepal: What a Register Search Cannot See
The real DOI trademark database limits in Nepal: unpublished pending applications, BS-only dates, extraction artefacts, and the conflicts no register shows.
A clean search result feels like a clearance, and in Nepal that feeling is dangerous. The DOI's published records are the deepest public dataset on Nepalese trademarks — and they still miss whole categories of conflict that decide real cases.
Knowing what the database cannot see is not a reason to skip the search. It is the reason to run the search properly and then layer the checks the database cannot do for you.
In short
DOI trademark database limits in Nepal come in three layers: the Industrial Property Bulletin publishes accepted applications only, so pending unpublished filings are invisible; the data arrives as PDFs in two scripts and two calendars, with extraction artefacts; and no register can weigh phonetic similarity, related goods across classes, unregistered market use or foreign marks never filed here. A clearance decision needs the database plus variant searching, transliteration checks and a market look.
What the DOI trademark database actually publishes
The Department of Industry's public record is the Industrial Property Bulletin: issue-by-issue PDFs listing each accepted application with the mark, applicant, class, goods or services, and dates in Bikram Sambat as well as AD. Issues reach back to 2006, and structured datasets built from them — the 71,000+ records behind this site's search tools — make the archive searchable in seconds.
That is the strength: publication-level visibility of what has been filed and accepted. The limits start where publication stops.
The window before publication
Applications that have been filed but not yet published do not appear anywhere. Examination takes months, and during that gap a filer can be ahead of you invisibly — you run a perfect search in March against a name someone filed in January that will not surface in the Bulletin until much later.
In a first-to-file system this is a structural risk, not a data flaw. It is managed by behaviour rather than searching: file early, because the search reserves nothing, and keep watching the Bulletin after your own filing so a later conflicting application is met inside the 90-day opposition window.
Data-quality limits: two scripts, two calendars, one PDF pipeline
- Extraction artefacts: letter-spacing glitches ('C H E M I C A L S') can break exact-match searching
- Device marks: the text field on a label mark may not match the artwork, so word-searching alone misses lookalike labels
- BS-dominant dates: publication dates — the ones that start opposition clocks — print in Bikram Sambat and must be converted carefully
- Class and goods text: suffers the same artefacts, so a class filter is a screen, not a verdict
The conflicts no database can weigh
Some of the most common refusals and oppositions turn on judgements, not records. Phonetic similarity — PH versus F, C versus K, Y versus I — is argued, not looked up. Related goods sit in different classes, so a same-class filter both over- and under-includes. Unregistered prior use in the market leaves no register trace at all. And a foreign mark never filed in Nepal is invisible here even if every consumer knows it.
| Risk source | Visible in DOI data? | Where it gets checked |
|---|---|---|
| Registered / published marks | Yes | Register search, exact and variant |
| Pending unpublished applications | No | File early; watch the Bulletin after filing |
| Unregistered prior use | No | Market and trade check in your channel |
| Company / firm names | No — separate register | Company registrar search |
| Foreign marks not filed in Nepal | No | Owner's portfolio, if known |
Layering the check: from database to decision
Exact search
The proposed name as spelled, across both name fields the Bulletin carries. Exact hits are the high-certainty conflicts.
Phonetic and spelling variants
Search the variants you would be annoyed to be opposed by: sound-alikes, doubled letters, plural forms, Y/I endings.
Transliteration check
The same brand can appear as EVEREST, एभरेष्ट or a romanisation of it. Cross-script variants are where naive exact-match searching fails completely.
Class and goods read
Filter by class, then read the goods text: a same-class hit for overlapping goods is a conflict; one for unrelated goods may not be.
The human layer
Market presence, company names and common-sense similarity close the gap the database cannot.
Managing around DOI trademark database limits
Three habits convert the limits from blind spots into managed risk. File promptly once a name clears the screen — the unpublished window punishes waiting. Watch the Bulletin continuously after filing — the database's blind spot is exactly where an opposition right is won or lost. And treat the screen as a filter, not an opinion: its job is to tell you whether a name is obviously taken, obviously free, or worth a practitioner's hour. Most names are one of the first two; the third is where the limits matter.
Want this handled for your brand?
IP Watch monitors new trademark publications in Nepal and alerts you to potentially conflicting marks — with the context needed to review them.
This article is general information, not legal advice.