Running on your device · 0 bytes uploaded

# Mojibake detector and repair

Undo the classic UTF-8-read-as-windows-1252 garbling (é, “, €) — each repair re-encodes and verifies the bytes decode back to valid UTF-8, so nothing is guessed. Double-mojibake repairs pass by pass.
Garbled textDetect and repair

- One family of mis-decode is covered: UTF-8 bytes shown as windows-1252/Latin-1 — the overwhelmingly common web case. Other pairs (Shift-JIS, KOI8-R mixes) are refused honestly, not half-fixed.
- When bytes do not decode back to valid UTF-8, the text is reported untouched — a partial repair would invent characters.
- Prevention beats repair: declare charset=utf-8 and the whole class disappears.
Nothing in yet

Paste, drop, or type to begin. Everything stays on this device.

paste garbled text

## Next door in Text

- [Unicode inspector](/text/unicode)
- [Base64](/encoding/base64)
[All text tools](/text)

---

Canonical HTML: https://nutter.tools/text/mojibake
Markdown version: https://nutter.tools/text/mojibake/index.md
Plain-text version: https://nutter.tools/text/mojibake/index.txt
Agent index: https://nutter.tools/llms.txt