Guides ·

BIP39 Word List: All 10 Languages Explained (2026)

When you create a cryptocurrency wallet, the 12 or 24 words you write down are not pulled from an arbitrary dictionary. They come from a carefully curated word list defined by the BIP39 standard. Most people encounter the English list --- 2,048 words from "abandon" to "zoo" --- but the specification actually defines word lists in 10 languages. Each list was designed with the same constraints: exactly 2,048 words, no ambiguity, and a structure that minimizes transcription errors.

This guide covers every official BIP39 word list, explains the design principles behind each one, and discusses the critical compatibility implications of choosing a non-English list.

Why 10 Languages?

BIP39 was designed to make cryptocurrency backups accessible to a global audience. The original specification (published in 2013) included only English, but contributors from around the world proposed additional word lists in the following years. The goal was to allow native speakers of major languages to write down seed phrases using words they could easily recognize, spell, and verify.

Each word list must satisfy the same core constraints:

  1. Exactly 2,048 words. This is non-negotiable. Each word represents 11 bits of data (2^11 = 2,048), and the mathematics of the entropy-to-mnemonic conversion depends on this exact number.
  2. The first 4 characters of each word must be unique. This allows users (and hardware wallets with small screens) to identify any word from just its first four characters.
  3. Words should be common and easily recognizable. Obscure vocabulary increases the chance of transcription errors.
  4. Visually and phonetically similar words should be avoided. "Woman" and "women" would be a bad pair. So would "accept" and "except."

These constraints are harder to satisfy in some languages than others, which is why some word lists took years to finalize while others were adopted quickly.

The 10 official BIP39 word lists are: English, Japanese, Korean, Chinese Simplified, Chinese Traditional, Spanish, French, Italian, Czech, and Portuguese.

English: The Universal Default

The English word list is the de facto standard. Virtually every wallet --- hardware, software, and web-based --- supports it. If you use Bitcoin, Ethereum, Solana, or any other BIP39-compatible chain, the English list is what your wallet will use unless you explicitly select otherwise.

Design Properties

  • 2,048 words from "abandon" to "zoo."
  • No word shorter than 3 characters. The shortest words are "add," "age," "ago," "aim," "air," and similar common terms.
  • First 4 characters are unique. "Abou" uniquely identifies "about." "Abso" uniquely identifies "absorb." This is why some hardware wallets only require you to type the first four letters.
  • No homophones or near-homophones. Words like "know" and "no," or "their" and "there," are excluded.
  • Alphabetically sorted. Binary search over the list is trivial.

The English list's widespread support makes it the safest choice for long-term storage. If you need to recover your wallet 10 years from now using wallet software that does not exist yet, the English word list is overwhelmingly likely to be supported. For a full technical explanation of how these words encode your entropy, see BIP39 Explained.

CJK Word Lists: Japanese, Korean, Chinese

The CJK (Chinese, Japanese, Korean) word lists present unique challenges and opportunities. These languages use logographic or syllabic scripts rather than the Latin alphabet, which changes the design constraints fundamentally.

Japanese

The Japanese word list uses hiragana characters exclusively. Each word is a common Japanese term written in the hiragana syllabary, making it readable without knowledge of kanji.

  • Script: Hiragana only (no katakana, no kanji).
  • Separator: The specification recommends ideographic spaces (U+3000) rather than standard ASCII spaces between words. This is a common source of implementation bugs.
  • Uniqueness: Each word is unique in its first 4 hiragana characters. Since each hiragana character encodes more information than a Latin letter, this constraint is easier to satisfy.

The Japanese list was one of the earliest non-English lists to be proposed and has relatively broad wallet support, particularly among Japanese exchanges and wallet applications.

Korean

The Korean word list uses hangul --- the Korean phonetic writing system. Each word is a common Korean word written entirely in hangul.

  • Script: Hangul only.
  • Character properties: Korean hangul characters are composed of jamo (consonants and vowels) arranged in syllable blocks. The visual distinctiveness of these blocks helps reduce transcription errors.
  • Word selection: Common, everyday words were preferred. The list avoids Sino-Korean loanwords that might be unfamiliar to younger speakers.

Korean wallet support is more limited than English or Japanese. If you choose to use the Korean word list, verify that your preferred wallet software explicitly supports it before generating a seed phrase.

Chinese Simplified and Chinese Traditional

BIP39 defines two separate Chinese word lists: one for Simplified Chinese (used in mainland China, Singapore, and Malaysia) and one for Traditional Chinese (used in Taiwan, Hong Kong, and Macau).

  • Script: Each list uses its respective character set exclusively.
  • Character count: Chinese characters encode more information per character than Latin letters. Each word in the Chinese lists is typically 1--2 characters, making the seed phrase visually more compact.
  • Overlap: Some words appear in both lists with identical meanings. Others exist in one but not the other, reflecting the natural divergence between simplified and traditional character sets.
  • Uniqueness constraint: The first-4-character rule applies differently here because many words are only 1--2 characters. The specification requires that no two words share the same first character in single-character words.

The Chinese lists are well-supported in wallets popular in Chinese-speaking regions but may not be available in Western wallet software.

European Languages: Spanish, French, Italian, Czech, Portuguese

Five European languages round out the official BIP39 word lists. Each uses the Latin alphabet with language-specific diacritical marks (accents, tildes, cedillas).

Spanish

The Spanish word list includes common Spanish vocabulary, with diacritical marks included as part of the word. For example, words with acute accents (like "arbol" vs. "arbol" with accent) are distinct entries. Wallet implementations must handle Unicode correctly to avoid treating accented characters as their unaccented equivalents.

  • Diacritics: Accented characters (a, e, i, o, u) are included and significant.
  • Regional neutrality: The list avoids words that are specific to a single Spanish-speaking country. "Computadora" (Latin America) and "ordenador" (Spain) might both be avoided in favor of a universally understood alternative.

French

The French word list follows similar principles to the Spanish list, with attention to accented characters (e, e, a, e, u, c, etc.).

  • Diacritics: Fully significant. "Resume" and "resume" with accent are different strings.
  • Ligatures: The list avoids characters like ae and oe that might not render consistently across systems.

Italian

The Italian list is relatively straightforward among the European lists. Italian uses fewer diacritical marks than French or Spanish, and the vocabulary pool is large enough to find 2,048 unambiguous words without difficulty.

Czech

The Czech word list is notable because Czech uses an extensive set of diacritical marks: haceks (caron), carons, and acute accents. Characters like c, r, z, s, and others are considered distinct from their base forms.

  • Diacritics: Essential and load-bearing. Stripping diacritics would create ambiguity.
  • Implementation risk: Software that normalizes Unicode aggressively (stripping accents, converting to ASCII) will corrupt Czech seed phrases. This is the single biggest risk with non-English word lists.

Portuguese

The Portuguese list was the most recent addition to the official BIP39 specification. It uses diacritical marks common in Portuguese (tildes, acute accents, circumflex accents) and attempts to be neutral between Brazilian and European Portuguese.

Compatibility Impact: Stick With One Language

This is the most important practical takeaway from this entire guide: your choice of word list is a permanent commitment. A seed phrase generated with the Japanese word list cannot be recovered using the English word list. The words are not translations of each other --- they are entirely different mappings from the same binary entropy to different word sequences.

The Mapping Is Language-Specific

When your wallet generates a seed phrase, it takes 128 or 256 bits of entropy, computes a checksum, splits the result into 11-bit segments, and looks up each segment in the word list. Word index 0 in the English list is "abandon." Word index 0 in the Japanese list is a different word entirely. The binary data is the same; the human-readable representation is different.

This means:

  • You cannot mix words from different lists.
  • You cannot "translate" a seed phrase by looking up each word's index and finding the corresponding word in another language.
  • Actually, you can do that translation mechanically (look up the index of each English word, then find the word at the same index in the Japanese list), and the result will be a valid Japanese-language seed phrase that derives the same wallet. But no standard wallet software provides this feature, and doing it manually is error-prone.

Wallet Support Is Not Universal

While English is supported by every BIP39-compatible wallet, non-English lists have varying levels of support:

Language Wallet Support Level
English Universal
Japanese Broad (major wallets)
Chinese Simplified Moderate (CJK-region wallets)
Chinese Traditional Moderate (CJK-region wallets)
Korean Limited
Spanish Moderate
French Moderate
Italian Limited to Moderate
Czech Limited
Portuguese Limited

If you generate a seed phrase with the Czech word list and then need to recover it 5 years later on a hardware wallet that only supports English, you are stuck. The only workaround is to find software that supports Czech BIP39, recover the wallet, and transfer funds to a new wallet generated with the English list.

Unicode Normalization

Non-English word lists introduce Unicode normalization as a potential point of failure. The BIP39 specification requires NFKD (Normalization Form Compatibility Decomposition) normalization before the mnemonic is fed into PBKDF2. This is especially relevant for:

  • Ideographic spaces (Japanese): The space character between Japanese words should be U+3000, not ASCII space (U+0020). Some wallets use one, some use the other. If the normalization step handles this incorrectly, the derived seed will be different.
  • Accented characters (French, Spanish, Czech, Portuguese): A character like "e" with accent can be represented as a single Unicode codepoint or as "e" + combining acute accent. NFKD normalization is supposed to resolve this, but buggy implementations exist.
  • Chinese character variants: Some Chinese characters have multiple Unicode representations. The word list uses canonical forms, but user input might use variant forms.

These issues are rare, but when they occur, they are catastrophic. Your seed phrase looks correct, but it derives a different wallet because of an invisible Unicode difference.

The Recommendation

Unless you have a compelling reason to use a non-English word list, use English. It has universal wallet support, no diacritical mark complications, and no Unicode normalization pitfalls. The English word list has been battle-tested across millions of wallets for over a decade.

If you do choose a non-English list, document which list you used alongside your seed phrase backup. Store a copy of the word list itself (all 2,048 words) with your backup. And test recovery on at least two different wallets before trusting the setup with significant funds.

SafeSeed's BIP39 Word Lookup

SafeSeed's seed phrase generators --- for Bitcoin, Ethereum, and Solana --- use the English BIP39 word list by default, ensuring maximum compatibility across wallets and hardware devices.

The tools display each generated word alongside its index in the BIP39 word list (0--2,047). This transparency lets you verify that the word-to-index mapping is correct and cross-reference against the official BIP39 English word list published in the Bitcoin BIP repository.

All generation happens entirely in your browser using the Web Crypto API for entropy sourcing. No word list lookups, no seed data, and no private keys ever leave your device. You can disconnect from the internet before generating and work in complete isolation. For a deeper understanding of why this matters, see Online Seed Generator Safety.

The BIP39 word list system is one of the most thoughtfully designed components of the cryptocurrency stack. It transforms raw binary entropy into something a human can write on a piece of paper, read back without error, and use to recover a wallet years later on hardware that has not been invented yet. The availability of 10 language options makes this system accessible worldwide. But that accessibility comes with a responsibility: choose your language deliberately, understand the compatibility implications, and keep your backup strategy aligned with the wallet ecosystem you rely on.