BOM

Last updated April 23, 2026

BOM stands for “Byte Order Mark”, the Unicode character U+FEFF placed at the very start of a text file to signal the encoding and, in wide encodings, the byte order. A UTF-8 BOM looks like the byte sequence EF BB BF. A UTF-16 LE BOM is FF FE. A UTF-16 BE BOM is FE FF.

Why it exists

In the multi-byte encoding era (UTF-16, UTF-32), you needed to know whether the file was little-endian or big-endian to decode it correctly. The BOM solved this: if the first two bytes are FF FE, the file is little-endian UTF-16; if FE FF, big-endian. A decoder reading a file could inspect those first bytes and know what to do.

Why it is awkward for UTF-8

UTF-8 has no byte-order ambiguity — it is defined as a specific sequence of 1-to-4-byte codes with no endianness. So a UTF-8 BOM is not really a byte-order mark; it is just a magic signature that says “this file is UTF-8”. Some tools write one. Others don’t. Some readers tolerate it. Others don’t.

The result: the UTF-8 BOM is an optional, invisible, often-invisible-in-editors character at the start of a file that may or may not break parsers.

Where it causes problems

Strict JSON parsers — the JSON spec does not allow a BOM before the first structural character. Some parsers error with Unexpected token in JSON at position 0 (that is an invisible BOM you cannot see). Others silently accept.
Bash scripts — a BOM at the start of #!/bin/bash makes the shebang line invalid, and the script fails to execute.
Source code files — some compilers trip on a BOM.
JSONL / NDJSON — a BOM at the start breaks parsing of the first record.

How to get rid of it

# Strip BOM from a single file using sed
sed -i '1s/^\xEF\xBB\xBF//' file.json

# Using tr (pipe-friendly but loses the file)
tr -d '\357\273\277' < in.json > out.json

# Using PowerShell
(Get-Content -Raw in.json) -replace "^", "" | Set-Content -Encoding utf8NoBOM out.json

How to prevent it

In VS Code: click the encoding label in the status bar → “Save with Encoding” → choose “UTF-8” (not “UTF-8 with BOM”).
In Notepad (Windows 10+): select “UTF-8” from the encoding dropdown instead of “UTF-8 with BOM”.
In PowerShell: use Out-File -Encoding utf8NoBOM instead of utf8.
In Git: configure .gitattributes with * text=auto eol=lf and use editors that default to UTF-8 without BOM.

A quick test

# First three bytes of the file:
head -c 3 file.json | od -c

# If you see "357 273 277", you have a UTF-8 BOM.
# Expected for clean UTF-8: just the first real character.

UTF-8 — the encoding that sometimes has an unnecessary BOM.
The guide The 10 most common JSON syntax errors lists BOMs as error #10.

Back to glossary Open JSON tool