A cheat sheet for your terminal.

A few small, copyable recipes for everyday engineering. These examples use harmless sample data and UTF-8 text.

macOS and Linux

Use printf to avoid the extra newline that echo usually adds. Command-line options vary between platforms.

# Encode text (macOS and Linux)
printf '%s' 'Hello, world!' | base64

# Decode on Linux (GNU coreutils)
printf '%s' 'SGVsbG8sIHdvcmxkIQ==' | base64 --decode

# Decode on macOS
printf '%s' 'SGVsbG8sIHdvcmxkIQ==' | base64 -D

# Encode a file without line wrapping on GNU/Linux
base64 -w 0 sample.bin

# Encode a file without line wrapping on macOS
base64 -i sample.bin -b 0

JavaScript in Node.js

const text = 'Hello, 世界 🌍';
const encoded = Buffer.from(text, 'utf8').toString('base64');
const decoded = Buffer.from(encoded, 'base64').toString('utf8');

// URL-safe alphabet; padding is omitted
const token = Buffer.from(text, 'utf8').toString('base64url');

Node’s decoder is permissive. If validating external input, check its alphabet and padding for your protocol before decoding.

JavaScript in a browser

btoa() expects one-byte characters; do not pass arbitrary Unicode directly to it. Convert text to UTF-8 bytes first. This example is for short strings.

const bytes = new TextEncoder().encode('Hello, 世界 🌍');
const encoded = btoa(Array.from(bytes, b => String.fromCharCode(b)).join(''));

const decodedBytes = Uint8Array.from(atob(encoded), c => c.charCodeAt(0));
const text = new TextDecoder('utf-8', { fatal: true }).decode(decodedBytes);

Python 3

import base64

raw = 'Hello, 世界 🌍'.encode('utf-8')
encoded = base64.b64encode(raw).decode('ascii')
decoded = base64.b64decode(encoded, validate=True).decode('utf-8')

url_safe = base64.urlsafe_b64encode(raw).decode('ascii')
# Remove '=' only when your protocol requires no padding.

Measure bytes, not characters

An emoji may use four UTF-8 bytes. Character counts, JavaScript string lengths, and byte counts can all differ. For padded Base64, calculate size using 4 × ceil(byteLength / 3). On this site the input counter reports UTF-8 bytes.

Reference documentation

See Node.js Buffer, Python’s base64 module, MDN btoa(), and GNU base64.