Fix broken umlauts and special characters

Prompt that finds the step where umlauts and special characters break and plans a fix and a repair. Paste it into an AI tool of your choice, which then finds the encoding error.

Topic: Apps & prototypes
Type: Prompt
Level: Intermediate
Time: about 15 min

Results depend on the tool, the model and what you enter.

Result preview

Example: written by our editors, not real AI output.

Symptom: "GrÃ¶ÃŸe" instead of "Größe" Cause: UTF-8 file read as Latin-1 during CSV import Path: shop export (UTF-8) → import script (Latin-1) → database (UTF-8) Fix: set encoding="utf-8" in the import Repair: only rows with "Ã" from the import date, on a copy first Check: sample word "Größe" in the import test data

Fill in

What you type stays in this browser window. We store and send nothing.

SymptomWhat do you see?

PathWhere does the text travel?

StackWhich technology?

Your prompt

Help me find and fix an encoding problem.
Symptom: [Symptom]
Path: [Path]
Stack: [Stack]
Approach: 1. From the broken characters, work out which encodings were probably mixed up, for example UTF-8 read as Latin-1. 2. List every step the text passes from input to output and the encoding each step uses or assumes. 3. Name the step where the conversion goes wrong and how to confirm it with one known sample word. 4. Propose the fix at the source and, separately, how to repair data that is already broken without damaging correct rows. 5. Suggest a check that catches the problem early in future.
Output: likely cause, path with encodings, fix, repair plan for existing data, preventive check.

Take particular care to follow one known word through every step and never repair the original data directly.

Before you answer, check: Are all instructions met? If key information is missing or a placeholder in square brackets is still in the text, ask instead of guessing. Mark assumptions as assumptions.

Help me find and fix an encoding problem.
Symptom: [Symptom]
Path: [Path]
Stack: [Stack]
Approach: 1. From the broken characters, work out which encodings were probably mixed up, for example UTF-8 read as Latin-1. 2. List every step the text passes from input to output and the encoding each step uses or assumes. 3. Name the step where the conversion goes wrong and how to confirm it with one known sample word. 4. Propose the fix at the source and, separately, how to repair data that is already broken without damaging correct rows. 5. Suggest a check that catches the problem early in future.
Output: likely cause, path with encodings, fix, repair plan for existing data, preventive check.

Take particular care to follow one known word through every step and never repair the original data directly.

Before you answer, check: Are all instructions met? If key information is missing or a placeholder in square brackets is still in the text, ask instead of guessing. Mark assumptions as assumptions.

This is an instruction for an AI, not the result itself.

Where this prompt works

These AI tools are a good fit for this prompt:

How to use it

Fill in the fields or tap a suggestion. Leave out names, addresses and other personal data.
Copy the finished prompt into an AI tool of your choice.
Pick one word with several special characters, such as "Größe", and follow it through every step; the first place where it looks wrong is the culprit.

Why does this work?

Encoding errors almost always come from one step that reads text in a different encoding than it was written in. Following one known word through the chain finds that step faster than guessing.

Tips

Back up the affected tables before any repair, and run the repair on a copy first.

You may use our prompts freely, including at work. Please do not republish them as your own collection. We give no warranty for AI results; check them yourself. Terms of use

Fix broken umlauts and special characters

Result preview

Fill in

Your prompt

Where this prompt works

How to use it

Why does this work?

Tips

Also a good fit

Choose pagination for an API list New

Write an API reference people can use right away New

Plan a throttled data backfill New