Why the download has a BOM. When you double-click a CSV, Excel never asks which encoding to use — it falls back to the system default and shows UTF-8 text as mojibake. A three-byte invisible signature at the start of the file (the BOM: EF BB BF) tells Excel “this is UTF-8”, so the accented letters are right on the first open. That is why the CSV you download starts with one. If the file is going into a database or an older system that does not strip BOMs, use the .xlsx export instead.
The browser's plain toUpperCase() turns “istanbul” into ISTANBUL, with a dotless I. In Turkish the correct result is İSTANBUL, because i and ı are different letters. This tool applies the Turkish rules — type something and watch the difference:
If your spreadsheet is full of ç ı ÅŸ é Ãœ or Ýstanbul, þube, the data is not lost — the file is simply being read with the wrong encoding. This free tool detects what the bytes really are (UTF-8, Windows-1254, ISO-8859-9), repairs the broken characters and hands back an Excel-ready UTF-8 CSV with a BOM or an .xlsx file. No sign-up, no watermark, no upload: everything runs in your browser and your file never leaves your device.
That privacy is not a nice-to-have. The spreadsheets that need this fix are usually customer lists, payroll exports and account ledgers — personal data that, under the Turkish KVKK and the European GDPR, should not be handed to an unknown website just to fix a few letters. Here not a single byte is sent anywhere, and closing the tab leaves nothing behind.
Drop your CSV, TXT or Excel file into the panel on the left — or paste the broken text straight into the box.
The encoding is detected and reported in plain words. In the before/after tables, broken cells are red and repaired ones are green.
Optionally apply the phone, IBAN, plate and Turkish case cleaners in the second tab, then download UTF-8 CSV or XLSX.
| What you see | What it should be | What happened |
|---|---|---|
ç ı ÅŸ ÄŸ ö Ãœ | ç ı ş ğ ö Ü | A UTF-8 file was read as Windows-1252/1254 — “double-encoded UTF-8”, by far the most common CSV encoding problem. |
Ý þ ð Þ Ð ý | İ ş ğ Ş Ğ ı | A Windows-1254 file was read as ISO-8859-1 (Western European). Classic in exports from older accounting and ERP systems. |
? or ��� | ç ğ ı İ ş | A legacy single-byte file was assumed to be UTF-8 and the undecodable bytes collapsed. Recoverable only while the original file still exists. |
’ “  | ’ “ (nbsp) | Smart quotes and non-breaking spaces pasted from Word went through the same double-encoding chain. |
The tool recognises all four cases separately and writes on screen which one it found. If there is nothing to fix it says the file is already clean instead of quietly changing something.
The second tab tidies the repaired table column by column: it normalises phone numbers to +90 5XX XXX XX XX (0532…, 90532… and 0090532… all land on the same result), strips and regroups IBANs into blocks of four, upper-cases licence plates, trims stray spaces and collapses repeated ones. The case conversions follow Turkish rules, not the default ones: toUpperCase() renders “istanbul” as ISTANBUL, while this tool produces İSTANBUL, and it lower-cases “IĞDIR” to ığdır rather than iğdır. The live example box on the page lets you type any word and see both results next to each other.
A CSV file carries no record of the encoding it was written with. If the file was saved as UTF-8 but Excel opens it with the system default (Windows-1252 in the US and Western Europe, Windows-1254 in Turkey), ç turns into ç, ı into ı and ş into ÅŸ. The opposite mistake, reading a legacy single-byte file as UTF-8, produces question marks or black diamonds instead. This tool detects the real encoding of the bytes and rebuilds the correct letters.
No. Reading, detection, repair and export all happen inside your browser tab, so the file never leaves your device. That matters because the files people need to fix are usually customer lists, payroll sheets and account ledgers — personal data that should not be handed to a random website. Nothing is stored and nothing is sent.
A BOM (Byte Order Mark) is a three-byte invisible signature at the very start of the file: EF BB BF. When you double-click a CSV, Excel does not ask which encoding to use — it falls back to the system default. If it finds a BOM it treats the file as UTF-8 and shows every accented character correctly. That is why the CSV you download starts with one. If you are feeding the file into a database or an older system that does not strip BOMs, download the XLSX instead.
Both are single-byte Turkish encodings and the letters sit at exactly the same positions; they only differ in the 0x80-0x9F range, which holds punctuation such as curly quotes and dashes. The WHATWG Encoding Standard that browsers implement maps the ISO-8859-9 label onto the Windows-1254 table anyway, so both options produce identical output here. The choice is kept for clarity, not because it changes anything.
Yes. Files up to 25 MB are parsed completely, so every column cleaner and the XLSX export are available. Larger files, up to roughly 400 MB, are processed in streaming mode: the encoding repair covers the whole file, the preview stops at the first 50 rows, and column cleaners plus Excel export are switched off. The tool says so on screen rather than pretending otherwise.