PRIVATE BY DEFAULT. INSTANT BY DESIGN.

CSV Mojibake & Encoding Repair

Repair reversible UTF-8 / Windows-1252 corruption, strip NUL bytes and export UTF-8 CSV.

Source CSV
0 characters

Pasted text is already decoded; source encoding applies to files only. Browser ISO-8859-1 decoding follows Windows-1252. Ambiguous corruption and replacement characters cannot always be recovered; review the result.

Your CSV files stay in this browser.

Before & after: Fix CSV Encoding

BEFORE · INPUT

name,note
José,“Hello”
Muñoz,’

AFTER · OUTPUT

name,note
José,“Hello”
Muñoz,’

How to use Fix CSV Encoding

01

Add CSV data

Paste CSV or drop local files. Joining requires a primary table and lookup table.

02

Choose options

Select match columns and join rules, chunk limits or source encoding and line endings.

03

Process locally

Run the task in a cancellable browser worker. Review the preview and result summary.

04

Download your result

Download joined or repaired UTF-8 CSV, or a ZIP of chunks with repeated headers.

Real-world workflows

Data Analysts

Use Fix CSV Encoding to prepare fix csv encoding results for repeatable reporting, exports, and data quality checks.

Software Engineers

Use Fix CSV Encoding to transform source content for integration fixtures, debugging, or code review directly in the browser.

E-Commerce Managers

Use Fix CSV Encoding to adapt product catalog exports and order spreadsheets for downstream inventory workflows.

Technical edge cases & frequently asked questions

Do my CSV files leave this browser?

No. Parsing, joins, repair and ZIP generation run in local browser workers. No source is uploaded.

How are duplicate and blank join keys handled?

Duplicate right keys expand to all matching rows. Blank keys do not match. Full joins copy unmatched right keys into the left match column. Matching trims whitespace and ignores case by default.

What are the practical limits?

Combined decoded sources have a 64 MiB cap; joins cap results at 1,000,000 rows and splitting caps output at 1,000 files. Device capacity varies; no under-50-ms guarantee. Quoted multiline fields are preserved and every chunk repeats its header.

Can all broken encodings be repaired?

No. Files use strict UTF-8 detection with Windows-1252 fallback; ISO-8859-1 is ambiguous and browser decoding maps it to Windows-1252. Reversible mojibake can be repaired, but lost replacement characters cannot be recovered. Review punctuation before export.

Technical specifications
EncodingUTF-8 (text)
ProcessingLocal JavaScript / browser Canvas
Max recommended file size50 MB+ may work; browser memory and tool complexity determine capacity. Start smaller.
Server retention0 bytes — no server uploads

Related Formatting Utilities