The tool
What the Llama strip HTML tool does
Paste HTML, get the readable text. The tool removes script blocks together with the code inside them, removes style blocks together with the CSS inside them, then removes every remaining tag and leaves the text those tags wrapped.
Removing the contents of script and style blocks is worth calling out, because the usual way this goes wrong is a stripper that deletes the tags and leaves the JavaScript sitting in your output. That does not happen here. Paste a page with a script block and the code is gone, not orphaned.
It runs in your browser. There is no upload, no account, no length limit and no cap on how often you run it.
Entities
Which HTML entities this tool decodes
Four, and only four. This is the detail that decides whether the output is usable for your text, so it is stated as a list rather than buried in a sentence about entity support.
- Decoded
- becomes a space, & becomes an ampersand, < becomes a left angle bracket, > becomes a right angle bracket.
- Not decoded
- " and ' stay as written, so quotation marks and apostrophes in some sources come back as entity text.
- Also not decoded
- ©, —, ’ and every numeric entity such as ’ pass through unchanged.
- Double-encoded text
- Because the four replacements run in sequence, something written as &lt; is decoded twice and arrives as a bare angle bracket rather than as <.
How to
How to strip HTML from Llama text in three steps
No settings, no sign-in, and the original stays in the input box so you can compare.
- Step 1
- Paste the HTML. A fragment, a full page, or a chunk of a saved article all work, and there is no character limit.
- Step 2
- Press strip HTML. The text appears in the output box immediately.
- Step 3
- Read the output before copying, especially if the source had entities or unusual attributes.
Structure
What happens to line breaks and page titles
Line breaks survive. The tool removes tags and nothing else, so wherever your source had a newline the output still has one. A heading followed by a paragraph on the next line comes back as two lines. There is no option to collapse them, because the tool does not touch whitespace at all.
Head content is kept, which is the one behaviour most likely to surprise you. If you paste a complete HTML document, the text inside the title element is stripped of its tags like everything else, so the page title appears at the very start of your output, usually running straight into the first line of the body.
That is easy to fix once you expect it: delete the first fragment, or paste only the part of the page you actually wanted.
Limits
Where the Llama strip HTML tool gets it wrong
This is a pattern-based stripper, not a full HTML parser, and there are two inputs that reliably defeat it. Both are rare in ordinary content and both are worth knowing about.
- An angle bracket inside an attribute
- A tag written as <a title="a > b" href="#"> ends the match early, so part of the attribute text leaks into your output.
- An angle bracket inside a comment
- An HTML comment containing a > breaks the same way and leaves the remainder of the comment visible.
- Malformed markup
- Unclosed tags and broken nesting are not repaired. What the pattern does not recognise as a tag is treated as text.
- The practical rule
- Ordinary article HTML is fine. Hand-written markup with comparison operators in attributes is where you should check the output line by line.
The usual pitfall
Why leftover script code is the classic failure
Guides to stripping HTML warn about the same thing every time: a naive pass removes the script tags but leaves everything between them, so the text you copy carries a few hundred lines of tracking code in the middle of the article.
Style blocks do the same thing with CSS. The visible result is a wall of selectors and colour values wedged between two paragraphs, which is obvious in a short document and very easy to miss in a long one.
Both are handled here before the general tag removal runs, so the contents go with the tags.
Uses
Why you would strip HTML out of Llama text at all
Text copied out of a rendered page, a CMS field or an email template often arrives carrying markup you did not ask for. Pasted into a plain text field, that markup shows up literally instead of formatting anything.
The other common reason is feeding a page to a model. Markup inflates the input, and the tags carry no meaning for the text you actually want summarised or rewritten. Stripping first leaves the content and drops the scaffolding.
The Llama in the name describes the text people usually arrive with. The operation itself is identical for every input, whatever produced it.
In practice
Who uses the Llama strip HTML tool, and when
Writers pulling copy out of a CMS field that stores markup. Support staff cleaning an email template down to its words. Anyone preparing a saved article as plain input before handing it to a model.
It is the right tool when the markup is disposable and the words are what you want. It is the wrong tool when the structure is the point, because nothing here converts tags into headings, lists or links.