🗜️ Run-Length Encoder / Decoder

Run-length encoding (RLE) compresses repeated characters into a count + character form (e.g. aaabbc → 3a2b1c). Simple, and reversible.

Mode

Encode Decode

Input

Count first, character second - always

The encoder walks the text and closes a run whenever the character changes, writing the length in front of the character. Single characters are not exempt: they are written as a run of one.

aaabbbccccd  ->  3a3b4c1d     11 characters become 8
hello        ->  1h1e2l1o      5 characters become 8

That second line is the whole story of RLE. It only shrinks data that actually repeats; on ordinary prose, where runs of two are rare, it roughly doubles the size. Encoding is safe on any input, and the Input Length and Output Length rows tell you immediately whether you gained anything.

Decoding, and where a round trip breaks

Never encode text containing digits. 1112 encodes to 3112, which the decoder reads as 311 copies of the character 2. The format has no way to tell a count from a digit in your data.
Keep it to one line. The decoder's pattern matches a count followed by any character except a line break, so a run of newlines is skipped silently and your paragraphs come back joined.
The decoder is strict about the rest. Anything that is not a count-plus-character pair leaves a gap, and you get an error rather than a mangled result.
Both directions run as you type; switching the radio button re-runs the current box, so you can paste the output back to check it.

This is a teaching format, not a compressor. To make a file smaller, use zip or gzip - they use LZ77 with Huffman coding, which finds repeated phrases rather than repeated single characters.

Frequently asked questions

What is run-length encoding used for?

Runs of identical bytes: fax transmission, BMP, PCX and TGA bitmaps, and the long white stretches of a scanned page. It is also a favourite first exercise in programming courses because the whole algorithm fits in ten lines.

Why is my encoded text longer than the original?

Because each character costs at least two characters once a count is added. Unless your text has long repeated stretches, RLE expands it - that is expected behaviour, not a fault.

How do I get my original text back?

Switch to Decode and paste the encoded string. As long as the original held no digits or line breaks, you get back exactly what you started with.