Word Replacer: Rules & Regex par Fortress

<p><strong>Word Replacer</strong> rewrites words and phrases on the pages you visit, using rules you write. Swap jargon for plain language, rename things you'd rather not read, fix a term a site always gets wrong, or translate recurring phrases — on every site, or only the ones you pick.</p>

<h3>Matching</h3>
<ul>
<li><strong>Literal or regular expression</strong> — plain text for everyday swaps, full regex with capture groups ($1, $&) when you need them.</li>
<li><strong>Smart case preservation</strong> — one rule handles every form. "dog → cat" turns DOG into CAT and Dog into Cat, without three separate rules.</li>
<li><strong>Whole-word matching</strong> that understands accents and non-Latin scripts. "naive" will not match inside "naiveté" — it uses Unicode boundaries rather than the usual ASCII-only word break.</li>
<li><strong>Case sensitivity</strong> per rule, not as a global setting.</li>
</ul>

<h3>Rules that don't fight each other</h3>
<p>Every rule is matched against the original text, and overlaps are resolved by priority. A rule turning "a" into "b" alongside one turning "b" into "c" leaves "a b" as "b c" — never "c c". Most replacers apply rules one after another and quietly produce the wrong answer here. Drag rules to set which wins.</p>

<h3>Scoping</h3>
<p>Limit any rule to specific sites or exclude it from them, with wildcard patterns like *.example.com. There's a global site blocklist, and a one-click pause for the site you're on.</p>

<h3>Import and export</h3>
<p>Move rules between machines or share a set with someone else. JSON and CSV both work, and importing shows you a preview — how many rules are new, how many are duplicates, what's invalid — before anything is written. Three merge strategies, and an undo that puts back exactly what you had.</p>

<h3>Careful with your pages</h3>
<ul>
<li>Never touches text you're typing — inputs, textareas and contenteditable regions are left alone.</li>
<li>Switching it off restores the page exactly. No reload needed.</li>
<li>Regex patterns are tested for catastrophic backtracking before they ever run on a page, with a time budget as a second line of defence.</li>
<li>Work is spread across idle time, so large pages don't freeze.</li>
<li>Keeps up with pages that load content as you scroll.</li>
</ul>

<h3>Privacy</h3>
<p>No network requests. No analytics, no telemetry, no account, no remote code. Your rules live in local browser storage and go nowhere else.</p>

<h3>Interface</h3>
<p>A full options page for managing rules, groups, import/export and preferences, plus a popup for quick toggles and adding a rule on the spot. Light and dark themes, and everything works from the keyboard — including reordering rules, so drag isn't the only way.</p>

<p><a href="https://github.com/DerMaxiAufGit/Word-Replacer-Extension">Source code on GitHub</a> — MIT licensed.</p>