Convert any file to a data URI in your browser, or decode one back into a file — with exact Base64 character counts and overhead figures, no uploads. Runs entirely in your browser — your numbers never leave this device.
No file encoded yet. The file is read on this device and turned into a data URI locally — the encoded text is exactly 4 characters for every 3 bytes, padding included.
| File size (bytes) | Base64 characters | Overhead |
|---|---|---|
| 3 | 4 | 33.3% |
| 8 (a PNG signature) | 12 | 50.0% |
| 300 | 400 | 33.3% |
| 1,024 | 1,368 | 33.6% |
| 1,048,576 (1 MiB) | 1,398,104 | 33.3% |
Base64 turns binary bytes into text that can travel anywhere text can — inside HTML, CSS or JSON — by re-packing the bytes: every group of 3 bytes (24 bits) is split into four 6-bit characters, with padding characters (=) completing a partial final group. That arithmetic fixes the cost exactly: the encoded form is always 4 × ceil(bytes ÷ 3) characters, about one third larger than the file. This tool reads your file in the browser and shows both numbers side by side — the file's exact byte count, the exact character count of its Base64 form, and the overhead percentage — so the size trade of embedding is a measured fact, not a rule of thumb you have to trust.
Encode mode produces the full data URI (the file's MIME type, the ;base64 flag, and the payload) in a copyable text area, plus ready-made CSS background and <img> snippets with copy buttons. Decode mode runs the other way: paste any data URI and the tool parses its MIME type and encoding, reports the decoded size, previews it when it is an image, and offers the reconstructed file as a download. The parser accepts plain (percent-encoded) data URIs as well as Base64 ones, and says so plainly when what you pasted is not a data URI at all.
Files never leave your device in either direction. Encoding reads the file with your browser's local file reader; decoding parses and rebuilds the file in memory. Nothing is uploaded, stored or logged — pasting a data URI here is exactly as private as pasting it into your own text editor, and the tool works offline.
No — in either mode. Encoding reads the file locally in your browser and decoding rebuilds the file in local memory; no file, data URI or fragment of one is sent to a server, stored or logged. Disconnect from the network and both modes still work, which is the check you can run yourself.
Because of the packing: 3 bytes become 4 characters, a ratio of exactly 4/3, so large files converge to 33.3% overhead. Small files can pay more — an 8-byte file needs 12 characters, +50% — because the last partial group is padded up to a full 4 characters. The tool shows your file's exact figures rather than the average.
For small, page-critical graphics — a tiny logo, an icon, a background pattern — where saving one network request matters more than the 33% size tax, and where the URI can be cached inside a CSS or HTML file you already ship. For photographs and anything large, a separate image file is smaller and cacheable on its own; embedding a 1 MiB photo adds roughly 350,000 characters to your page.
Scope. The developer tool encodes and decodes text, with URL-safe alphabets and a small file mode. This tool is built for files and images: it produces complete data URIs with the MIME prefix, reports exact byte and character counts, generates CSS and <img> snippets, previews decoded images and downloads the decoded file with the right extension.
Usually one of three things: the string does not start with data: (a bare Base64 blob has no MIME header — the developer Base64 tool handles those), the comma separating header from payload is missing, or the payload was truncated when copying, which breaks the final padded group. The tool names a non-data-URI when it sees one and refuses a payload it cannot decode rather than downloading a corrupt file.
More in Image Tools · Browse all tools