ts-base58

A direct TypeScript port of deemru/b58 — Bitcoin-alphabet Base58 encoding.

If you find this project useful, you can support this and further ports at ko-fi.com/scottmoore0.

Upstream provenance

This package is a TypeScript port of deemru/b58 (b58.c, b58.h) by Dmitrii Pichulin, MIT License.

The translated output is validated against canonical Bitcoin Base58 test vectors including the standard "Hello World!" → "2NEpo7TZRRrLZSi2U" reference plus encode-then-decode round-trips on arbitrary 32-byte inputs (simulates a private key).

Why this exists

Base58 is the encoding used for Bitcoin addresses, IPFS CIDs (v0), Solana public keys, Monero addresses, Stellar account IDs, Ripple, and many cryptocurrency / decentralised-system identifiers. The 58-character alphabet excludes 0, O, I, l to avoid visually-ambiguous characters when transcribing by hand. This is not base58check (which adds a 4-byte SHA-256-SHA-256 checksum) — for that, layer a SHA-256 (via ts-sha2) on top.

Install

npm install ts-base58

Usage

import { base58Encode, base58Decode } from 'ts-base58';

base58Encode(new TextEncoder().encode('Hello World!'));
// '2NEpo7TZRRrLZSi2U'

const decoded = base58Decode('2NEpo7TZRRrLZSi2U');
new TextDecoder().decode(decoded);
// 'Hello World!'

// Encoding a Bitcoin-style 32-byte private key
const privkey = new Uint8Array(32);
crypto.getRandomValues(privkey);
const wif = base58Encode(privkey);

API surface

• base58Encode(input: Uint8Array | string): string — encode to Base58.
• base58Decode(input: string): Uint8Array — decode from Base58.

Both empty-input and leading-zero (0x00... prefix) inputs are handled correctly — Base58 encodes leading zero bytes as leading '1' characters.

Reference values

The test suite asserts against canonical Bitcoin Base58 vectors:

| Bytes | Base58 | |---|---| | (empty) | (empty) | | 0x00 | 1 | | 0x00 0x00 | 11 | | 0x61 | 2g | | 0x62 0x62 0x62 | a3gV | | 0x63 0x63 0x63 | aPEr | | "Hello World!" | 2NEpo7TZRRrLZSi2U | | "The quick brown fox jumps over the lazy dog." | USm3fpXnKG5EUBx2ndxBDMPVciP5hGey2Jh4NDv6gmeo1LkMeiKrLJUUBk6Z |

Plus round-trip checks (32-byte simulated private key) and alphabet sanity (no 0, O, I, l characters in output).

Run:

npm test

Caveats

• Not base58check. This is plain Base58 — no checksum. For Bitcoin addresses use base58check by appending a 4-byte SHA-256-SHA-256 checksum before encoding (build it with ts-sha2).
• Not Flickr / Ripple alphabets. This is the Bitcoin alphabet: 123456789ABCDEFGHJKLMNPQRSTUVWXYZabcdefghijkmnopqrstuvwxyz. Other Base58 variants (Flickr swaps letter case order; Ripple uses a different ordering) need a different package or post-processing.
• Performance. Base58 encoding is O(n²) in the input length due to the carrying-into-arbitrary-precision-arithmetic nature of the algorithm. Fine for keys/hashes (typically 32 bytes); slower for very long inputs.

License

MIT. Original C by Dmitrii Pichulin under MIT.

See also

• ts-chibi-base64 — Base64 encoding (RFC 4648)
• ts-sha2 — SHA-256 (for building base58check)
• cpp-to-ts — the translator that produced this package