Gera Tools

Netstring Encoder/Decoder

Encode strings in Bernstein's length:data, netstring format

Wrap any string as a netstring — a byte count, a colon, the raw data, then a comma — exactly per DJB's spec. Decode netstrings back to text and validate the length prefix. Runs entirely in your browser. It runs free in your browser on Gera Tools, with nothing uploaded.

Last updated Source: Gera Tools

What is a netstring?

A netstring is a self-delimiting encoding invented by Daniel J. Bernstein. It is the byte length of the data in ASCII decimal, a colon, the raw data, then a trailing comma. For example the string hello becomes 5:hello, which is unambiguous to parse.

A netstring is a tiny, self-describing way to wrap a string so a parser always knows exactly where it ends. Invented by Daniel J. Bernstein, it is widely used in protocols like SCGI and in qmail. This tool encodes any text into a netstring and decodes netstrings back to their payload, all in your browser.

The netstring format

The complete format is described in three characters of grammar: length:data, — where length is the byte count of data expressed in ASCII decimal digits. That is the entire specification.

5:hello,        → the string "hello"  (5 bytes)
0:,             → empty string        (0 bytes)
11:hello world, → "hello world"       (11 bytes)

No escaping, no quoting, no reserved characters. Because the parser reads exactly length bytes after the colon, the payload can contain any byte value including :, ,, null, and binary data.

How encoding and decoding work

Encoding:

  1. Measure the input data in UTF-8 bytes using TextEncoder (not character count — a single emoji is 4 bytes).
  2. Write that count as ASCII decimal digits.
  3. Append :, the raw data bytes, then ,.

Decoding:

  1. Read ASCII digits until the first :.
  2. Parse the digits as the declared byte length.
  3. Slice exactly that many bytes as the payload.
  4. Require a , immediately after. If the comma is missing, the netstring is malformed.

Why bytes, not characters?

The length prefix counts bytes in the UTF-8 encoding, not Unicode codepoints or visible characters. This is critical when the payload contains multi-byte characters:

StringUTF-8 bytesEncoded netstring
hello55:hello,
café5 (é is 2 bytes)5:café,
😀44:😀,
(empty)00:,

Always count bytes, not characters, or the decoder will read the wrong number of bytes and corrupt subsequent messages in a stream.

Where netstrings are used

Netstrings appear in:

  • SCGI (Simple Common Gateway Interface): The web-server-to-application protocol used by nginx and others to pass request headers.
  • qmail and djbdns: Bernstein’s own mail and DNS tools use them extensively for IPC.
  • Length-prefixed messaging in custom protocols: Any situation where you want safe streaming of arbitrary payloads without a framing character that could appear in the data.

Streaming multiple netstrings

You can concatenate netstrings directly without a separator, since each one is self-delimiting:

5:hello,5:world,   → two strings: "hello" and "world"

A decoder reads the first length prefix, slices the payload, confirms the comma, then immediately starts the next number — making this suitable for framing messages over a TCP stream or in a file.

When decoding, a prefix that claims a length longer than the remaining data is flagged as an error, so truncated or corrupted streams are detected immediately rather than silently producing garbage.