When a name like René Müller shows up as René Müller, the GEDCOM file was read with the wrong character set: UTF-8 text was treated as Windows text (often called ANSI). You can repair it for free with our GEDCOM checker, which shows every garbled line next to what it should say and fixes only the lines it can convert back exactly.
Last checked: 7 October 2026. The checker is ours; what it does is described below exactly as it works.
How to recognise the problem
Garbled characters follow a pattern. An accented letter turns into two characters, usually starting with à or Â, and quotation marks turn into three starting with â€.
| Should be | Shows as | Should be | Shows as |
|---|---|---|---|
| é | é | ö | ö |
| è | è | ü | ü |
| ë | ë | ä | ä |
| ç | ç | ß | ß |
| ñ | ñ | ’ | ’ |
A small example from a GEDCOM file, before and after:
Broken:
1 NAME René /Müller/
2 PLAC Köln
Fixed:
1 NAME René /Müller/
2 PLAC Köln
Three different causes
- The garbled text is inside the file. A UTF-8 file was opened as Windows text and saved again. Every program now shows é, because that is what the file says.
- The file is fine, but the header names the wrong character set. The text is UTF-8, but the header says
1 CHAR ANSELor1 CHAR ANSI. A program that trusts the header shows é; one that looks at the bytes shows é. - The opposite: question marks or black diamonds. The file is in ANSEL or Windows text and the program reads it as UTF-8. Then you see René as Ren? or Ren�.
The GEDCOM 5.5.1 standard lists four character sets for the header (ANSEL, UTF-8, UNICODE and ASCII) and calls ANSEL the preferred one. Many Windows programs write ANSI anyway. GEDCOM 7.0 only allows UTF-8. That history is why files that travel between programs so often end up garbled.
How to fix it with the free checker, step by step
- Open the GEDCOM checker and choose your .ged file. It is read in your browser and never uploaded.
- At the top you see the character set: what the header says and what the text really is. If they differ, the checker reports that the header names the wrong character set (cause 2).
- Under Garbled characters you see how many lines are affected and, per line, the text as it is and as it would be repaired. Check a few examples.
- If they look right, tick Also repair the garbled characters and download the repaired file. It ends in -repaired.ged.
- Import the repaired file into your genealogy program, or open it in our GEDCOM viewer to see that the names are right.
What the checker does with each cause:
- Cause 1: it repairs the garbled pieces, but only when you tick the box and only where turning the characters back into bytes gives valid UTF-8. Anything else stays as it was: a strange character is better than a wrong letter.
- Cause 2: it reads the text as what it really is, and the repaired file says
1 CHAR UTF-8. - Cause 3: it reads ANSEL and Windows text correctly and writes the repaired file as UTF-8. Bytes that match no known character are counted and shown, but cannot be restored.
The repair never changes dates, places, notes or links. It also makes line endings uniform, removes empty lines and indentation, and adds a missing header or end marker.
Other ways to fix it
| Way | Good for | Watch out for |
|---|---|---|
| Our free GEDCOM checker | All three causes; shows each line before it changes anything; nothing to install | Garbled text it cannot convert back exactly stays as it is |
| Export again from the program that made the file | The best fix if the original is still correct in that program; choose UTF-8 if it asks | Does not help if the garbled text is already in the program's own data |
| Gramps (free desktop program) | When the header says ANSEL, Gramps asks on import which character set to use (ANSEL, ANSI, ASCII or UTF-8); its GEDCOM export is UTF-8 | Does not undo text that is already garbled inside the file |
| A text editor with an encoding setting, such as Notepad++ | Reading the file in another encoding without changing the bytes, or converting it to UTF-8 | Easy to make it worse: converting an already garbled file garbles it again; its manual does not mention ANSEL |
How to prevent it
- When your program asks for a character set on export, choose UTF-8.
- After exporting, open the file in the checker once: the header should say
1 CHAR UTF-8and the names under Do these names look right? should look right. - Do not open and save a GEDCOM file in a word processor or a text editor that guesses the encoding.
- Sending a file to someone with an older program? Ask what it reads. Older programs sometimes expect ANSEL; see how to convert an ANSEL GEDCOM file to UTF-8 for the other direction of the story.
Related: broken links, duplicate IDs and impossible dates, how to export a GEDCOM file and how to open a GEDCOM file without software.
Sources
- The GEDCOM Standard, release 5.5.1 (character sets, line length, cross-reference IDs, dates): gedcom.io/specifications/ged551.pdf
- FamilySearch GEDCOM 7.0 specification (UTF-8 only, no CONC, GEDZIP): gedcom.io/specifications/FamilySearchGEDCOMv7.html
- Gramps source code, GEDCOM import (encoding choice for files marked ANSEL): github.com/gramps-project/gramps/blob/maintenance/gramps60/gramps/plugins/importer/importgedcom.py
- Gramps source code, GEDCOM export (writes GEDCOM 5.5.1 with CHAR UTF-8): github.com/gramps-project/gramps/blob/maintenance/gramps60/gramps/plugins/export/exportgedcom.py
- Notepad++ User Manual, Encoding: npp-user-manual.org/docs/encoding