Hebrew transliteration tool
Convert Hebrew company names to Latin script for cross-script entity matching. Hebrew is a consonantal alphabet — the output reflects the consonants as written.
Output appears here…
How it works
Hebrew characters are converted to their Latin consonant equivalents — ב→b, ג→g, ד→d, מ→m, נ→n, and so on. Because standard business Hebrew omits vowel diacritics (nikud), the output is a consonantal skeleton rather than a fully-voweled romanization.
The pipeline then strips Hebrew legal suffixes (בע"מ for Ltd, בע"מ ציבורית for PLC), lowercases, and sorts tokens. When comparing two Hebrew-script records of the same company, both sides produce the same consonantal canonical — enabling reliable deduplication within Hebrew-script datasets.
Worked examples
| Input (Hebrew) | Company | Canonical |
|---|---|---|
| בנק הפועלים | Bank Hapoalim | bnk hfv lym |
| אלביט מערכות | Elbit Systems | albyt m rkhvt |
| בזק | Bezeq | bzk |
| רפאל | Rafael | rfal |
* Output reflects consonants as written — vowels are not present in standard Hebrew business text.
Embed this tool
<iframe src="https://ambect.com/tools/hebrew-transliteration" width="100%" height="420" style="border:none;border-radius:12px" title="Hebrew transliteration tool — Powered by Ambect" ></iframe>
Frequently asked questions
What is Hebrew transliteration?
Hebrew transliteration converts Hebrew script consonants to their Latin equivalents. Unlike Arabic, which also omits most vowels in standard writing, Hebrew uses a consonantal alphabet (abjad) where vowel diacritics (nikud) are almost never written in business contexts. The output is a consonantal skeleton of the name.
Why does the output look like consonants only?
Standard Hebrew writing omits most vowels. The word "bank" (בנק) is written as three consonants: B-N-K. The Ambect pipeline faithfully transliterates what is written — it does not attempt to guess missing vowels. This is consistent with how most automated matching systems handle Hebrew text.
Is the consonantal output still useful for entity matching?
Yes. When comparing two Hebrew-script records of the same company, both sides will produce the same consonantal canonical. The Ambect API also applies synonym mappings to link common variants, improving match rates even when exact consonant patterns differ slightly.
How do I match Hebrew and English versions of the same company?
Submit both through the Ambect normalize/company endpoint with the same country code. The pipeline normalizes each to a canonical token set. Where a synonym entry exists linking the Hebrew form to the English brand name, both sides will resolve to the same canonical.
Need to match Hebrew and English company records at scale? The Ambect API normalizes both to the same canonical form in under 5 ms.
Get API access free →