Detect and remove zero-width characters from your text
When you paste text from a PDF, email, or a web page, invisible characters sometimes hitchhike along with the content. Zero-width spaces, soft hyphens, direction markers, and other hidden Unicode characters can break search functionality, cause unexpected word breaks, interfere with data validation, or serve as fingerprints to track where your text came from. Detecting and removing these invisible characters keeps your text clean, your data safe, and your privacy intact.
What are zero-width characters
Zero-width characters are Unicode code points that occupy no visual space. Unlike regular spaces or punctuation, they leave no visible mark even when you enable "show whitespace" in a text editor. Common examples include the zero-width space (U+200B), which acts like a space but takes up zero pixels; the zero-width joiner (U+200D), used in ligatures; left-to-right and right-to-left marks (U+200E and U+200F), which control text direction; and the soft hyphen (U+00AD), which indicates where a word can break across lines.
Why invisible characters break your data
These characters corrupt downstream systems in silent, frustrating ways. A database search for "example" fails if the actual text contains "example" with an invisible zero-width space between the words. Validators reject email addresses and URLs containing hidden marks. Copy-paste operations produce malformed output. In data pipelines and imports, they corrupt structured records without any visible error message. More concerning, threat actors deliberately embed zero-width characters as fingerprints to track document leaks—each recipient gets a unique pattern, so when the text reappears publicly, the source can be identified.
How to detect invisible characters
Manual inspection fails because you cannot see what is not visible. The reliable approach is to analyze the raw Unicode data. Paste suspect text into the invisible character detector, which examines every code point and flags anything that is zero-width, control, or otherwise invisible. The output shows the character's Unicode name, code point, and category, so you know exactly what you are dealing with. This transparency is crucial when validating data from untrusted sources.
Understanding invisible character ranges
Zero-width characters cluster in specific Unicode ranges that you should recognize. The range U+200B–U+200F contains space and joiner variants. U+202A–U+202E covers bidirectional formatting controls. U+2060–U+2064 includes word joiners and other invisible punctuation. U+2066–U+2069 holds newer directional isolates used in complex scripts. The invisible text tool shows these ranges interactively, so you can experiment and understand which characters appear in different contexts and why they matter.
Privacy—your data stays in your browser
These tools run entirely in your browser, not on any server. Your text never travels across the network, no account is required, and the tools keep working even without an internet connection. This is essential when auditing sensitive content—passwords, proprietary code, confidential documents—because you control where the data goes and who sees it. Detection and cleaning happen on your device, and the results remain yours alone. You maintain full privacy and full control over every piece of data you analyze.
Cleaning and normalizing your text
Detection is the first step; cleanup is the second. Some invisible characters should be stripped outright because zero-width spaces almost never belong in clean text. Others require context-aware handling depending on your use case. The normalize Unicode tool applies standard Unicode normalization forms, collapsing variant representations into canonical ones. For maximum thoroughness, the remove non-ASCII tool strips everything outside the basic Latin alphabet, which guarantees no hidden Unicode survives the process.
Real-world scenarios where you need this
Import data from PDFs, scanned documents, or web content and invisible characters tag along for the ride. Accept user input in web forms or databases and you should validate early to catch hidden marks before they corrupt your records. Share documents within a team and suspect a leak: scan for zero-width fingerprints to identify the source. Maintain code repositories: hidden characters in source code cause silent bugs and unexplained behavior. Migrate legacy data between systems: clean before import to prevent corruption in the new environment. Receive text from external APIs or automated systems: normalize before processing to ensure consistency.