Base64 decode online

Base64 decoding reverses encoding: a stream of ASCII characters becomes the original bytes, then readable UTF-8 text. Each valid quartet of Base64 characters represents three output bytes before padding adjustments.

When you decode base64 online, you are translating transport-encoded text into what your editor would show — not performing URL decoding, HTML unescaping, or JWT signature verification. Those are separate layers.

This base64 decoder uses atob followed by TextDecoder with UTF-8, matching what most modern web stacks expect for text payloads. Binary image or PDF data should use our Base64 to Image tool for preview and download.

Invalid characters, wrong padding, or alphabet mismatches surface as clear errors rather than silent corruption.

Paste your Base64 string into the input panel — with or without line breaks. The decoded text appears live as you edit. SGVsbG8= decodes to Hello; longer strings with emoji expand after decoding to multi-byte UTF-8 characters.

Enable URL-safe mode when the source uses - and _ instead of + and /, common in JWT middle segments and some query parameters. The tool normalizes padding automatically.

Copy decoded text with one click or save as a .txt file. For round-trip checks, paste the result into our Base64 Encoder and confirm you get the original Base64 back.

Import a .txt or .b64 file when the string is too long to paste reliably from a log viewer.

JSON APIs sometimes return Base64-wrapped text fields. Decode here to read the plaintext before chaining to a JSON formatter or URL decoder if further escaping appears.

JWT middle segments are Base64URL-encoded JSON. Paste the segment between the dots, enable URL-safe mode, and read the payload — but never trust unsigned tokens for authorization. Use a proper JWT library for signature verification.

If decoded text looks like another Base64 string, you may be facing double-encoding. Decode once, inspect the result, and decode again only if it still matches the Base64 alphabet.

Garbled characters after decode often mean the original data was binary (image, PDF) rather than UTF-8 text. Try Base64 to Image or a file-oriented decoder instead.

Invalid padding: standard Base64 length should be a multiple of four. The tool adds = padding for URL-safe input automatically, but corrupted strings from truncated copy-paste need manual repair.

Whitespace and quotes: log exports and JSON often wrap strings in quotes or insert newlines every 76 characters. This decoder strips whitespace; remove surrounding JSON quotes before pasting.

Alphabet mismatch: standard Base64 uses + and /. URL-safe variants use - and _. Toggle URL-safe mode when decode fails on strings from JWTs or query parameters.

Wrong content type: decoding image or PDF Base64 here shows unreadable characters. Use Base64 to Image for visual preview or download.

URL percent-encoding (%48%65%6c%6c%6f) is not Base64 — use our URL Encoder/decoder for that layer. Hex byte dumps need the UTF-8 Decoder, not this page.

HTML entities (&) are markup escaping — use our HTML Escape tool. Chain decoders in order: Base64 first if the outer wrapper is Base64, then UTF-8 or URL as needed.

base 64 decode and base64 decode refer to the same RFC 4648 process. Spacing, hyphens in keywords, and casual spelling do not change the algorithm.

atob and TextDecoder run locally in your browser. API keys, JWT payloads, and customer data never upload to a server. Clear the textarea when finished on shared machines.

Pair with our Base64 Encoder for reverse conversion. Use Base64 to Image when the payload is a photo or diagram. Use Image to Base64 when you need to create encoded strings from files.

Treat decoded content from untrusted logs carefully — malicious payloads should be inspected in isolated environments.