What the UTF-16 Byte Converter helps you do
Use this page to inspect the exact bytes used by UTF-16 little-endian and big-endian text, add or omit a byte-order mark, or restore text from a hexadecimal byte sequence. The workspace keeps the source, settings, and result together so you can understand what changed instead of receiving an unexplained clipboard value. Start with a small known example before using unfamiliar data.
The letter A is 41 00 in UTF-16 LE and 00 41 in UTF-16 BE. With a BOM, those examples begin FF FE or FE FF respectively. The visible result remains available for selection, copying, and download. Nothing is copied or saved automatically, so you can stop and compare it with the source before taking the next step.
How the conversion works
Each JavaScript UTF-16 code unit becomes two bytes. Little-endian writes the low byte first; big-endian writes the high byte first. A BOM is FF FE for little-endian or FE FF for big-endian. Decoding checks the byte count, BOM, selected order, and surrogate validity before showing text.
Auto is safest when the input includes a BOM. If no BOM exists, choose LE or BE from information supplied by the file or protocol. When encoding, Auto uses LE and the separate BOM control states whether the marker is included. Inputs and outputs stay in the current browser tab during ordinary use. NexaCurrent does not upload the working value to perform this operation. Reset clears the entered value, settings, result, messages, and generated download from the workspace.
How to use UTF-16 Byte Converter
Enter or paste the smallest complete source value that represents your task. Choose the source format, direction, byte order, output family, or fixture type shown on this page. Run the primary action, read the summary, and inspect the complete visible result before using Copy or Download.
Encode a value containing both a basic Latin letter and an emoji, then decode the bytes with the same order and confirm the exact text returns. If the receiving application has its own specification, sample, or validator, use that as a second check. A successful result here confirms this page’s stated rules; it cannot promise that another system has selected the same convention.
Errors, boundaries, and recovery
Odd byte counts, non-hexadecimal characters, a selected order that conflicts with the BOM, missing BOM in Auto mode, and isolated surrogate units are rejected rather than guessed. An error clears an earlier result so an obsolete success cannot be mistaken for the corrected answer. Fix the first message, then run the operation again. Changing the input or a setting also removes stale copy feedback.
Text input is limited to 200,000 characters, generated quantity to 100 items, and pattern or schema expansion to the smaller limits shown beside those controls. These bounds keep one browser tab responsive. For a large file or batch, use a streaming command-line tool and preserve an original copy.
Limits and responsible use
Byte order cannot be inferred reliably from arbitrary marker-free bytes. This page handles UTF-16 text bytes, not complete files with headers, record layouts, or application-specific metadata.
Do not paste a real password, private key, access token, personal record, or production customer dataset when a harmless sample can prove the same behavior. Treat downloaded output as a new file, keep the source until the target system accepts it, and document the selected format whenever another person must reproduce the result.