URL / Base64 / HTML

Choose an operation and enter text. Ordinary inputs update directly; above 32,768 characters use Convert (Ctrl/⌘+Enter in the editor). Swapping applies the inverse operation to the previous output; it is disabled if that output exceeds the input limit. Text is always shown as plain text, never executed as HTML.

URL component mode uses encodeURIComponent/decodeURIComponent: + stays a literal plus when decoding. Form-value mode uses application/x-www-form-urlencoded rules: spaces become +, and + decodes to a space. These modes encode one value, not an entire URL or form.

Base64 uses UTF-8, preserving a leading BOM. Decoding accepts standard or URL-safe alphabet, omitted padding and ASCII whitespace, but rejects mixed alphabets, bad padding and nonzero pad bits. Bytes must form valid UTF-8; this is not a binary-file decoder. Empty input is valid. HTML escaping replaces &, angle brackets and quotes; decoding supports HTML named/numeric character references in text context, including legacy missing semicolons and numeric replacements. It does not strip markup, decode recursively or sanitize HTML.

Image mode defaults to unchanged original bytes, retaining animation/metadata. PNG/JPEG/WebP conversion may resize down without enlarging; it produces a static image and may remove metadata. JPEG composites transparency onto white. Only supported actual export formats are accepted; there is no silent PNG fallback. The caption shows actual format and dimensions.

Limits: 1 MiB UTF-8 text input, 8 MiB image file, 16 million decoded pixels, resize side 64–8192 px. Browser decoding occurs before the pixel check and codec support varies. Text, operation, format and exact maximum-side input stay locally in this browser, including unfinished or invalid text. Files, results, Data URIs and the active mode tab are not saved; nothing is uploaded. Large restored text waits for Convert; images always require Create Data URI. Browser storage limits still apply.