⌗ Tool I
Is your trademark available for registration? Let's see in three steps.
Turkish-tuned Soundex and Levenshtein matching against the TÜRKPATENT and WIPO Global Brand Database baseline. Free result, GDPR/KVKK-compliant PDF report.
Why is a trademark pre-search essential?
A trademark registration is the only document that legally anchors the years of effort the owner has invested. Before obtaining it, the first task is to check whether an identical or highly similar mark has already been registered in the same class. Under Turkish IP Code (SMK) 6769 art. 5 (absolute refusal) and art. 6 (relative refusal), a conflicting record may lead to rejection of your application — and the application fee is non-refundable.
Three-layered matching
This tool runs four checks simultaneously:
- Exact: Turkish characters normalised (ş→s, ğ→g, ı→i, ç→c, ö→o, ü→u) for a direct match.
- Spelling similarity: Levenshtein edit distance — for short marks 1, medium 2, long marks 80%+ similarity is considered critical.
- Phonetic similarity: A Turkish-adapted Soundex — catches marks written differently but sounding alike, e.g. "Asil"/"Asyl" or "Karaca"/"Karadja".
- Class overlap: Substring or close-stem records in the same Nice class.
Limitations of this tool
At the moment this tool runs on a seed dataset of known Turkish trademarks; it is an honest baseline but not a full pre-search. A full pre-search must cover:
- The TÜRKPATENT trademark portal (
portal.turkpatent.gov.tr) - WIPO Global Brand Database, including international registrations — branddb.wipo.int
- EUIPO eSearch Plus (EU trademarks)
- Firms not yet registered but actively using a mark (under Court of Cassation "priority of use" doctrine)
For this reason, however favourable our result looks, your agent's full pre-search before filing is always required. The agent also evaluates absolute refusal grounds (art. 5) — for example lack of distinctiveness or genericness — which an automated tool cannot fully detect.
AI transparency
Soundex and Levenshtein are classical algorithms — deterministic mathematics, not AI. No step of this tool relies on an LLM's answer; once we move to live TÜRKPATENT data in Phase 9, the same transparency is preserved: the tool's output never replaces the agent's judgement — it accelerates the agent's process.
⌗ FAQ
Frequently Asked Questions
- Yes. Complete three steps and we send a privacy-compliant report to your inbox. We require verification for more than three queries in 24 hours; for unlimited searches, contact an agent.
- Soundex is a phonetic algorithm that reduces a word's sound to a 4-character code; we use a variant adapted to Turkish sounds (ş/s, ç/c, ğ/g, ı/i). Levenshtein counts the number of letter edits needed between two words — even 1–2 edits in short marks indicate a serious confusion risk (SMK 6769 art. 6 relative refusal).
- At present this tool runs on a seed dataset of known Turkish trademarks — full TÜRKPATENT live portal + WIPO Global Brand Database integration will follow in Phase 9. Even now, a 'low risk' result does not replace your agent's full pre-search before filing.
- No. Absolute refusal grounds under SMK 6769 art. 5 (lack of distinctiveness, genericness, etc.) may lie outside a pre-search. No one guarantees registration; your agent's role is to run the process correctly.
- Active retention is 24 months under KVKK / GDPR-aligned policy. After that, automatic anonymisation. Newsletter subscription is a separate consent — you can unsubscribe from the link in every email.
- Talk to an agent via WhatsApp or phone. Medium risk typically means there is spelling or phonetic similarity within your class; adding distinctive visual or word elements may reduce the risk. Your agent will give a concrete opinion based on Court of Cassation case law.