fci4096 — Chinese Idiom Mnemonic Wordlist (成语助记词)

A BIP39-inspired mnemonic wordlist and seed-derivation library built on 4096 standard Chinese four-character idioms (成语). Offline, local-only, no_std + WASM compatible.

中文说明见下方 中文简介 章节。

Overview

fci4096 is a mnemonic system inspired by BIP39, replacing the English word list with 4096 standard four-character Chinese idioms (chengyu). Each idiom encodes 12 bits of information (2^12 = 4096), providing higher information density than BIP39's 2048-word list (11 bits/word) for the same number of words.

Key advantages:

• Offline & local-only — zero network requests, zero configuration, no cloud services
• no_std compatible — runs on bare-metal embedded devices and hardware wallets
• WASM ready — browser-based key derivation via the wasm feature
• Low latency — O(log N) idiom lookup via compile-time binary search array
• Lightweight — core dependency stack is sha2 + pbkdf2 + hmac

Differences from BIP39

Feature	BIP39	fci4096
Wordlist size	2048	4096
Bits per word	11	12
Entropy (12 items)	128 bits	128 bits
Checksum bits	4 bits	16 bits
Wordlist language	English	Chinese idioms

Encoding Scheme

With a 4096-word list, each idiom encodes 12 bits. The checksum is 1/8 of the entropy length:

Word Count	Entropy	Checksum	Total Bits
12	128	16	144
15	160	20	180
18	192	24	216
21	224	28	252
24	256	32	288

The checksum is the leading N bits of SHA-256(entropy).

Installation

[dependencies]
fci4096 = "0.1"

Feature Flags

Feature	Default	Description
std	Yes	Standard library support
rand	Yes	Random number generation (mnemonic generation)
serde	No	Serialize/Deserialize support
wasm	No	WASM bindings

Usage

Generate a Mnemonic

use fci4096::{generate, IdiomMnemonicSize};

let mnemonic = generate(IdiomMnemonicSize::Idioms12).unwrap();
println!("{}", mnemonic.phrase());
// e.g. 哀而不伤 按部就班 草木皆兵 ...

From Entropy

use fci4096::from_entropy;

let entropy = [0x00u8; 16]; // 128 bits of entropy
let mnemonic = from_entropy(&entropy).unwrap();

Parse and Validate

use fci4096::{from_phrase, validate};

let phrase = "哀而不伤 按部就班 草木皆兵 ...";
if validate(phrase) {
    let mnemonic = from_phrase(phrase).unwrap();
    let entropy = mnemonic.to_entropy().unwrap();
}

Generate a Seed

let seed = mnemonic.to_seed("passphrase");
// Returns [u8; 64], using PBKDF2-SHA512 with 4096 iterations

CLI Examples

# Generate a 12-word mnemonic
cargo run --example generate 12

# Recover the seed
cargo run --example recover "idiom1 idiom2 idiom3 ..." [passphrase]

Use Cases

• Cryptocurrency wallets — Chinese-localized mnemonic generation and recovery
• Hardware wallets — no_std embedded key derivation with O(log N) idiom lookup
• Web applications — WASM-based browser key generation without server round-trips
• Encrypted backup systems — Human-readable idiom phrases for offline key backup
• Identity / DID systems — Deterministic 64-byte seed derivation from memorable phrases

Performance

Metric	Value
Idiom lookup	O(log N) binary search (≈ 12 comparisons)
Index storage	2 bytes/idiom (Vec<u16>)
Seed derivation	PBKDF2-HMAC-SHA512, 4096 iterations
Dependencies	sha2 + pbkdf2 + hmac + unicode-normalization
no_std	Yes (bare-metal, embedded, hardware wallets)
WASM	Yes (wasm32 target with getrandom/js)

Security

• Entropy source: Uses the getrandom crate for OS-level cryptographically secure random numbers
• Checksum: 128-bit entropy with 16-bit checksum (BIP39 uses only 4 bits), providing stronger error detection
• Seed derivation: PBKDF2-SHA512, 4096 iterations, passphrase is NFKD-normalized
• Wordlist: 4096 standard four-character idioms, sorted by pinyin, systematically reviewed to remove derogatory idioms

Documentation

• docs.rs — Full API documentation with examples
• CHANGELOG.md — Version history

中文简介

一套借鉴 BIP39 规范、基于中文四字成语构建的密钥助记词方案，词库选用 4096 条规范标准成语。方案面向中文使用场景打造，大幅提升助记词记忆、读写与备份的本土易用性，实现自主可控的本地化助记词体系。

每条成语承载 12 比特熵信息（2¹² = 4096）；在助记词条数相同的前提下，相较原生 BIP39 的 2048 词词表（单词仅承载 11 比特）拥有更高信息密度，同等数量助记词可容纳更多随机熵，安全冗余更强。校验和长度为原始熵长度的 1/8。

核心特性：

• 离线本地运行 — 无网络请求，零配置，无第三方云服务
• no_std 兼容 — 支持裸机嵌入式设备和硬件钱包
• WASM 支持 — 通过 wasm feature 实现浏览器端密钥派生
• 低延迟 — 编译期生成的二分查找数组，O(log N) 成语查找
• 轻量级 — 核心依赖仅 sha2 + pbkdf2 + hmac

适用场景： 加密货币钱包（中文本地化）、硬件钱包、Web 应用、加密备份系统、身份/DID 系统。

License

MIT