Convert UTF-8 to Base64

All three use the same 64-character alphabet for the first 62 positions (A-Z a-z 0-9). They differ at positions 62 and 63: Standard uses + /, URL-safe uses - _, MIME uses + / but wraps lines at 76 chars with CRLF for fitting into 80-column email transports. URL-safe is the only variant whose output is safe to paste directly into URL paths/queries and filenames.

Many simple Base64 implementations call btoa(text) directly. That throws on any character above U+00FF because btoa only accepts binary strings (Latin-1). This tool calls new TextEncoder().encode(text) first to get the UTF-8 byte array, then passes those bytes through btoa – so any Unicode character including emoji round-trips correctly.

Base64 maps every 3 input bytes (24 bits) to 4 output characters (each carrying 6 bits). The grid shows each 3-byte block with its hex / decimal bytes and the resulting 4-character quartet – useful for debugging encoding output by hand or showing students how Base64 works.

Exactly 4/3 = 1.33× for the encoded bytes vs source bytes, plus padding (up to 2 chars). For ASCII source text it’s 4/3 ≈ 1.33× the input character count. For UTF-8 text with many multi-byte chars (emoji = 4 bytes each), the overhead per character can look much higher: 🌍 (1 input char, 4 UTF-8 bytes) → 8 Base64 chars including padding = 8× the input char count.

Yes – toggle “Include = padding” off. Some systems (JWT, modern URL-safe Base64 spec, OAuth) explicitly forbid padding. The decoder accepts both padded and unpadded input, adding padding back during decode for atob.

The decoder uses TextDecoder('utf-8', {fatal: true}), so if the decoded bytes are not valid UTF-8 (e.g., you Base64-encoded a JPEG or some arbitrary binary), it throws “Decoded bytes are not valid UTF-8”. For arbitrary binary, you need a different tool that doesn’t try to interpret the bytes as text.

RFC 2045 requires CRLF (rn) every 76 chars. This tool emits CRLF for MIME variant and LF (n) for the custom-wrap option. The decoder strips all whitespace including line breaks before processing.

No. It’s an encoding scheme designed to represent binary data in 7-bit ASCII. Anyone with the encoded string can decode it instantly. Don’t confuse Base64 with cryptographic security – if you need confidentiality, encrypt first (AES, RSA), then Base64-encode the ciphertext.

Yes for text/plain (what this tool defaults to). For images, you’d want a file converter that reads the bytes directly. The data URI here just wraps the Base64 in the standard prefix so you can paste into an HTML src attribute and the browser will decode.

No. Everything runs client-side via TextEncoder, btoa, atob, and TextDecoder. About 18 KB of code.