Base85 to file.

Base85, or ASCII85, writes four bytes as five characters drawn from the 85 printable ASCII characters from ! to u, with z standing for four zero bytes. Paul Rutter's btoa program introduced the idea in 1985; Adobe adopted it in PostScript Level 2 and in PDF (the ASCII85Decode filter), which is where most of it lives today.

Four bytes are read as one 32-bit number and written as five base-85 digits, each offset by 33 to land on a printable character. The output is a quarter larger than the input, the least of the common encodings. A final group of one to three bytes is written as two to four characters; a group of four zeros is a single z.

PDF streams and PostScript files, where binary images and fonts are embedded as text; Git's binary patches, with its own alphabet; ZeroMQ's Z85 variant for keys; and anywhere the extra density over Base64 is worth the unfriendly characters.

Denser than Base64 (25% against 33%) and that is its only advantage: the alphabet has quotes, slashes and brackets, so it needs escaping in a URL, a JSON string or a shell, and there are several incompatible variants. Use it where a format asks for it; otherwise Base64.

There are several Base85s. This one is ASCII85, as PDF and PostScript write it, optionally between <~ and ~>. Git's binary patches and ZeroMQ's Z85 use the same density with different alphabets and no z shorthand, so their text won't decode here and ASCII85 won't decode there.

No. Base85 is an encoding, not encryption: it changes how bytes are written, not who can read them, and anyone can decode it with no key. Encrypt first (AES, age, a library's sealed box) when the content needs protecting, then encode the ciphertext if it has to travel as text.

Paste the Base85 into the input, or drop a text file that holds it. The tool decodes the bytes, says what they look like and how many there are, and shows the image when they are one. Check the file name, which is guessed from the bytes, and press Download.

From the first bytes. Most formats start with a signature, 89 50 4E 47 for PNG, %PDF for PDF, PK for ZIP, and the tool knows the common images, documents, archives, fonts and media, and tells text from the rest. It is a guess for the file name only; if you know what the bytes are, rename the file and it downloads under that name.

Usually the Base85 was cut short or carries something extra. Copy the whole value, including its end; leave out quotes and commas from the surrounding code. If the status line says the bytes look like text when you expected an image, the value wasn't the file's Base85 to begin with.

Nothing leaves your browser. The text is read into memory on your machine and the encoding is computed there, in plain JavaScript; the page makes no request with it. Load the page, go offline and it keeps working.

No. The "Base85 to file" tool is free and opens without a sign-in, an account or an install.

Nothing. The "Base85 to file" tool holds your input in the page's memory, so closing or reloading the tab clears it, and it saves nothing in this browser.

Yes: you can rename it before downloading. The default name is decoded plus the extension the format suggests, such as decoded.png.

When the bytes start like a PNG, JPEG, GIF, WebP, AVIF, ICO, BMP or SVG image, the page shows a preview before you download it. Other types show their size and the name of the format they start like.

The decoder takes the characters "!" to "u" and "z" for four zero bytes, with or without the <~ ~> delimiters, ignoring spaces and line breaks.

Closest to the "Base85 to file" tool: File to Base85, Base85 decoder, Base85 encoder, Base64 to file. Each opens on its own page, free and in the browser.

The "Base85 to file" tool is one of 32 in its group. The group's page, /tools/encoding, lists them all.

The tool is free. Its interface is built from shadcn.io's own components (Button, Input, Slider and the rest of the registry under /ui), which come in React, Vue, Svelte and Solid. Pro adds the blocks and templates for the product around it.