File Encoding Converter

No. Your file is never sent to our servers - all encoding detection, conversion, BOM handling, and preview rendering happens directly in your browser on your device using JavaScript and the browser-native TextDecoder / TextEncoder APIs. There are no network requests at all.

30 encodings across every major writing system: Unicode (UTF-8, UTF-16 LE, UTF-16 BE), ASCII, Western European (ISO-8859-1, ISO-8859-15, Windows-1252), Central European (ISO-8859-2, Windows-1250), South European (ISO-8859-3), North European (ISO-8859-4), Cyrillic (ISO-8859-5, Windows-1251, KOI8-R, IBM866), Arabic (ISO-8859-6, Windows-1256), Greek (ISO-8859-7, Windows-1253), Hebrew (ISO-8859-8, Windows-1255), Turkish (Windows-1254), Baltic (Windows-1257), Japanese (Shift_JIS, EUC-JP, ISO-2022-JP), Korean (EUC-KR), Simplified Chinese (GB2312, GBK), and Traditional Chinese (Big5). Both reading (decoding) and writing (encoding) are fully supported for every one of them.

The detector runs a hand-written priority pipeline: BOM detection first (highest confidence), then null-byte sniffing for UTF-16 LE / BE, then UTF-8 multi-byte sequence validation, then ASCII detection, then ISO-2022-JP escape-sequence scan, then byte-range scoring for the five CJK encodings (Shift_JIS / EUC-JP / EUC-KR / GBK / Big5), then frequency heuristics for Cyrillic (KOI8-R vs Windows-1251 vs ISO-8859-5), Greek (Windows-1253), Arabic (Windows-1256), Hebrew (Windows-1255), and Central European (Windows-1250), and finally an ISO-8859-15 vs Windows-1252 fallback based on the presence of the euro sign byte.

Confidence is High when a Byte Order Mark is present (definitive proof of encoding), Medium when valid UTF-8 multi-byte sequences are found (almost always correct), and Low when only single-byte heuristics matched. Single-byte encodings like ISO-8859-x and Windows-125x have overlapping byte ranges so a small file in one of them is genuinely ambiguous - check the source preview and override the source encoding manually if it looks wrong.

Those are unmappable characters - your text contains a character that exists in the source encoding but has no equivalent in the target encoding. For example, converting a UTF-8 file containing emoji to ISO-8859-1 will replace the emoji with "?" because ISO-8859-1 only has 256 code points. To preserve every character, pick a Unicode target like UTF-8 or UTF-16.

When the target is UTF-8 or UTF-16, a BOM control appears with three modes. Auto includes a BOM for UTF-16 (where it indicates byte order) and strips it for UTF-8 (where it is optional and breaks some tools). Include forces a BOM into the output. Strip removes any BOM from the output. Use Strip for PHP / Bash / shell scripts that break with BOM, and Include for CSV files you want Excel to open as UTF-8.

Only text-based files are accepted - 35+ extensions including .txt, .csv, .tsv, .log, source code (.js, .ts, .py, .java, .c, .cpp, .html, .css, .php, .rb, .go, .rs), data formats (.json, .xml, .yaml, .yml, .ini, .env, .toml, .conf, .properties), documents (.md, .rst, .tex, .rtf), .sql, subtitle files (.srt, .sub, .ass, .vtt), and .svg. Binary files like images, PDFs, ZIPs, and Office documents are rejected with an error. Maximum file size is 50 MB to keep the browser tab responsive.

No. This tool intentionally does not use localStorage - your file lives only in browser memory while the tab is open. Closing the tab, refreshing the page, or clicking Clear All wipes everything. Nothing is ever written to disk except the file you explicitly download with Convert & Download.

Supports UTF-8, UTF-16 LE/BE, ASCII, ISO-8859-1 / -2 / -3 / -4 / -5 / -6 / -7 / -8 / -15, Windows-1250 / 1251 / 1252 / 1253 / 1254 / 1255 / 1256 / 1257, KOI8-R, IBM866, Shift_JIS, EUC-JP, ISO-2022-JP, EUC-KR, GB2312, GBK, and Big5. Both reading (decoding) and writing (encoding) are fully supported for every one of them - this is not a UTF-8-only converter pretending to support other formats.

The detector runs a hand-written priority pipeline: BOM detection first (highest confidence), then null-byte sniffing for UTF-16 LE / BE, then UTF-8 multi-byte sequence validation, then ASCII detection, then ISO-2022-JP escape-sequence scan, then byte-range scoring for the five CJK encodings, then frequency heuristics for Cyrillic / Greek / Arabic / Hebrew / Central European, and finally an ISO-8859-15 vs Windows-1252 fallback based on the presence of the euro sign byte.

Files starting with the UTF-8 BOM (EF BB BF), the UTF-16 LE BOM (FF FE), or the UTF-16 BE BOM (FE FF) are detected with High confidence on the very first three bytes. The detection panel shows "BOM detected" so you know the result is definitive.

Files without a BOM are scanned byte-by-byte and validated against the UTF-8 specification: 2-byte sequences must start with 110xxxxx and have one continuation byte, 3-byte with 1110xxxx and two continuation bytes, 4-byte with 11110xxx and three. If every multi-byte sequence is well-formed and at least one exists, the file is reported as UTF-8 with Medium confidence.

Five different CJK encodings (Shift_JIS, EUC-JP, EUC-KR, GBK / GB2312, Big5) are scored by counting valid lead-byte / trail-byte pairs in the file. The encoding with the highest score wins, provided the score crosses a minimum threshold to filter noise. ISO-2022-JP is detected separately by scanning for the ESC ($ escape sequence.

For non-CJK files, byte-frequency heuristics distinguish KOI8-R vs Windows-1251 (different Cyrillic letter ranges), pick up ISO-8859-5 Cyrillic, Windows-1253 Greek, Windows-1256 Arabic, and Windows-1255 Hebrew based on the dominant byte ranges typical of each script.

Three levels with explicit reasoning: High ("BOM detected") for files with a Byte Order Mark, Medium ("valid UTF-8 byte sequences") for valid UTF-8 without BOM, and Low ("fallback heuristic") for single-byte encodings where the result is a best-effort guess. The confidence label appears in the detection details panel so you know how much to trust the result.

After upload (or after clicking Detect Encoding manually), a panel shows the detected encoding name, confidence level with reason, BOM presence (yes / no), and exact file size. Useful when you just want to identify an unknown file's encoding without converting it.

Single-byte encodings use a reverse character-to-byte map built from the browser's TextDecoder. Multi-byte CJK encodings use two-byte reverse lookup tables seeded by iterating valid lead/trail byte pairs through TextDecoder. UTF-16 LE/BE are written byte by byte with manual surrogate-pair handling. Unmappable characters (e.g. emoji in ISO-8859-1) are replaced with '?'.

When the target is UTF-8 or UTF-16, a BOM mode toggle appears. Auto includes a BOM for UTF-16 (where it indicates byte order) and strips it for UTF-8 (where it's optional and breaks some tools). Include forces a BOM. Strip removes any BOM. Use Strip for PHP / Bash scripts that break with BOM, Include for CSV files you want Excel to open as UTF-8.

The source area shows the first 500 characters of your file decoded with the currently selected source encoding. If the preview is garbled, change the Source Encoding dropdown and click Refresh preview to re-decode and visually verify until the text looks right.

The output preview re-renders automatically every time you change the source encoding, target encoding, or BOM mode (via a useEffect). No extra Generate or Preview Output button - the converted text appears as soon as you tweak any setting, so you can iterate fast.

Drag a text file from your file explorer directly onto the input area. The drop zone highlights while you drag. Files are validated for type and size on drop just like clicked uploads, with clear error messages for rejected files.

Only text-based files are accepted - 35+ extensions are whitelisted (.txt, .csv, .tsv, .log, source code, web files, data formats, subtitles, .svg). Binary files like images, PDFs, ZIPs, and Office documents are rejected before they're read, with a clear error message naming the unsupported extension. Files over 50 MB are also rejected to keep the browser tab responsive.

When you click Convert & Download, the converted file is saved with the original filename - not a generic "converted.txt". This means a UTF-8-to-Windows-1252 round trip of report.csv saves back as report.csv, ready to drop into the same workflow that produced the input.

Unlike most tools on this site, the File Encoding Converter intentionally does not persist anything to localStorage. Your file lives only in browser memory while the tab is open. Closing the tab, refreshing, or clicking Clear All wipes everything - the only output is the file you explicitly download.