“Accented payee names import as garbled characters”
When a payee named Café Müller arrives in your register with a stray capital A before every accent, nothing is corrupt — the bytes are intact and being read with the wrong alphabet. Bank files declare which alphabet they use, and importers do not always believe them.
Symptoms
Accented letters appear as two or three strange characters each A stray "Ã" or "Â" precedes otherwise-correct letters Some payees look fine while others do not, depending on which letters they contain
What actually causes it
UTF-8 bytes read as Windows-1252
This is the classic "é" pattern: the file is UTF-8, where é is two bytes, but the importer reads each byte separately using the older single-byte alphabet. One character becomes two.
A header that misdeclares the encoding
OFX 1.x headers state ENCODING and CHARSET. Banks sometimes emit UTF-8 content under a header declaring plain ASCII or 1252, and a compliant importer follows the header rather than the bytes.
A conversion step in the middle
Opening and re-saving the file in an editor or spreadsheet with a different default encoding rewrites the bytes, which is why files often break after being "checked".
How to fix it
- Drop the file into the fixer below: it reports when a header declares ASCII while the content contains accented characters, which pinpoints this exact mismatch.
- Prefer re-downloading over patching — a fresh export from the bank usually has consistent bytes and header.
- If you must edit the file, use an editor that lets you choose the encoding explicitly on both open and save, and keep it the same as it arrived.
- The fixer deliberately does not transliterate accented names, because guessing which alphabet was intended can quietly rename people.
Related import problems
Skip the manual edit — fix it automatically
The fixer flags the encoding mismatch rather than silently rewriting names, and downloads the repaired file in the same encoding it arrived in. Free, no signup, no upload.
Open the fixer