A browser extension that checks the page you are looking at for things that should never have gone live: dummy text, unfilled template tags, values like undefined printed by code, garbled characters and broken images. One click, on the page you choose, entirely on your device.

Who it is for

Anyone who publishes web pages: developers before a release, editors after a content migration, shop owners after changing a theme, agencies handing a site to a client, testers working through a checklist. Leftovers are easy to miss because the page "looks finished" at a glance: a price that reads NaN €, a greeting that says Hello {{ name }}, a footer with [Company Name], or umlauts that turned into Ã¼ after an import. Search engines and customers notice them anyway.

What it finds

Filler copy: Lorem ipsum and other dummy text, including German filler text, and phrases such as Your headline goes here.
Unfilled template tags: {{ … }}, {% … %}, <%= … %>, ${…}, [[…]], %%NAME%%, {FIRST_NAME} and unformatted message arguments such as {count} or %s.
Placeholders to fill in: [Your Name], [Phone Number], [Insert logo], [Ihr Name], runs of xxx and TBD.
Values printed by code: undefined, NaN, [object Object], Invalid Date, Infinity and a lone null. undefined in an ordinary sentence counts as low severity.
Garbled characters: UTF-8 text read in the wrong encoding (GrÃ¼ÃŸe instead of "Grüße"), shown together with the text it probably should be; replacement characters (�) and HTML entities shown as text (&).
Broken and placeholder images: images that failed to load, have no address, or whose file name says "placeholder".
Development leftovers: links and images pointing at localhost, private network addresses, file:// or staging hosts; TODO and FIXME notes; test data such as test@test.com.
Search and sharing texts: a default page title such as "Document" or "Untitled", and dummy text in the meta description or link-preview tags.

How to use it

Open the page you want to check and let it load completely.
Click the Leftover Text Finder button in the toolbar. The check runs immediately.
Read the findings, most serious first. Show scrolls to one finding and flashes it; Highlight on page outlines every finding in orange, and Remove highlight restores the page.
Click Copy report to paste a plain-text list into a ticket or an email.

What the levels mean

Level	Meaning
High	Almost certainly a mistake every visitor can see: dummy text, an unfilled template tag or placeholder, a value printed by code, garbled characters, a link to a development address.
Medium	Probably a leftover: a broken or placeholder image, a developer note, test data, a default page title, an unreadable character.
Low	Worth a look but often intended: links to "#" and the word `undefined` in a sentence. Hidden unless you ask for them.

What it deliberately ignores

Code samples (code, pre, kbd, samp), so documentation that shows template syntax is not flagged.
Hidden content such as closed menus and inactive tabs; it checks what visitors can see.
Text you are editing in a field or an editable area.
Links to local addresses when the page itself is on a local address, so it works during development too.

Limits

The rules are heuristics: read each finding before you change anything. A page about programming will show undefined in prose, which is why that case is low severity.
It checks the main page, not embedded frames from other sites, and only what has loaded so far.
Browsers do not let extensions run on internal pages such as settings, the extension store or the new-tab page.
It reports; it never changes the page's content, fixes anything or sends anything.

Privacy and permissions

The extension collects no data. It has no servers, no analytics and no account, and it stores nothing. It runs only when you click its button, only on the tab you clicked it on, and the report stays in the extension window until you close it.

Permission	Why it is needed
`activeTab`	Gives temporary access to the current tab only after you click the toolbar button. The extension cannot see any other tab or any page you have not clicked it on.
`scripting`	Runs the checker, which is included in the extension, in that tab to read the visible text, images and links, and to add or remove the highlight.

It requests no access to your browsing history, no host permissions and no background access. The full privacy policy is part of the site's privacy policy (Deutsch).

Looking for text that is on the page but hidden from people? Hidden Text Scanner finds concealed text, invisible characters and instructions aimed at AI assistants.

Support

Questions, or a finding that should not be one? Write to hello@oakenbit.com.

Leftover Text Finder