A public word filter and an edited source record do different jobs. If you want selected words masked for readers while keeping the original post or comment available to editors, use a display filter and test it against the output your site actually publishes.
This walkthrough uses Content Censor Free 3.2.3. The optional scanning step needs Pro. Content Censor doesn’t replace comment moderation, reporting, permissions or a security system.
Start with one harmless test word
Install and activate Content Censor Free, then open Content Censor → Settings. Replace the example list with a harmless sample term such as orange. Use the pipe delimiter if you add a second term, for example orange|pear.
Entries shorter than three characters are ignored. Keep your initial list narrow so you can identify an unwanted match before applying a large set of rules.
Choose how the word should match
Keep strict filtering enabled for the first test. Whole-word matching lets orange match that word without intentionally matching it inside a longer word. Keep case-insensitive matching if the rule should cover Orange as well.
Embedded matching is available when you need it, but a short fragment can affect unrelated words and names. Test those rules separately.
Limit the public output
Enable only the output area you want to test, such as main post content. Content Censor also has controls for supported titles, excerpts, comments and comment-author names, tags and tag clouds.
If legitimate content must remain unfiltered on a particular post or page, add its numeric ID to the exclusion list. Exclusions aren’t a substitute for reviewing the overall word list.
Pick a readable replacement
Choose the asterisk replacement and the option that masks every letter for a straightforward first test. Other rendering choices retain the first letter or the first and last letters. A blank replacement removes the matched word from view and may make a sentence harder to read.
Save the settings. Open a representative public page while signed out and check both the intended match and nearby words. Then open the original post in the editor: the source text should still contain the original word.
Review existing matches with Pro
If you use Pro, open the Scan Content tab in Settings and choose Scan Now. The scanner reviews published content and approved comments. Version 3.2.3 works through large comment threads in bounded batches and merges repeated matches for the same post.
Use the results as a review list. The scanner doesn’t change stored text, find every piece of dynamically assembled output or make moderation decisions.
If the result isn’t what you expected
Check the delimiter, matching mode, case setting, selected output areas and excluded IDs. Clear page caches and test while signed out. If markup changes unexpectedly, review the Ignore HTML setting and any broad terms in the list.
For support, copy the privacy-safe summary from Content Censor Home and add an anonymised example of the term and sentence. Don’t include passwords, licence keys or a database export.
The complete reference explains each setting. The product page compares the Free controls with Pro scanning and integration tools.