Text chunker
Compare fixed, sentence and Markdown splitting on your own text, with every chunk drawn to scale.
/chunk
Free · no account · nothing uploaded
HTML text, link and table extractor
Give it a page’s source and it returns what you actually wanted: the readable text without the navigation, every link with its anchor text, and any tables as structured rows.
HTML extractor
Say it the way you would say it to a person. Your words are matched to one of 18 tools by readable rules, and the tool runs right here in your browser. No upload, no account.
This tool runs in your browser and needs JavaScript turned on.
esc
How to use the HTML extractor
- Open the page in your browser, choose View Source, and copy the HTML.
- Paste it into the message box, or save it as an .html file and drop it into “Your files”.
- Ask for what you need: “extract the text and links” or “get the tables from this page”.
- Download the readable text, a CSV of every link, and a CSV for each table on the page.
Things you can type: extract the text and links get the tables from this page
Ninety percent of a page is not the page
Navigation, cookie banners, footers, related-article rails, three layers of wrapper divs. Getting the article out of that is a small, tedious job that stands between you and every other thing you wanted to do with the content — chunk it, count it, quote it, cite it.
This does that step and hands back three things you can use directly.
What comes back
- Readable text, with markup and boilerplate stripped, ready to chunk or summarise.
- Every link on the page with its anchor text — the raw material for a link audit or a crawl frontier.
- Each table as its own CSV file, so a table on a page can go straight into the CSV tools.
It parses HTML rather than pattern-matching it
The extraction runs through a real DOM parser, so malformed markup, nested tables and attributes containing angle brackets behave the way a browser would treat them. Regular expressions over HTML get this wrong in ways that are hard to notice and harder to debug.
The tool takes its parser by injection — the browser supplies its own, and the headless test suite supplies one too, which is why the extraction logic can be tested without a browser at all.
Fetching versus extracting
This page cannot fetch a URL for you: a browser will not let one site read another’s HTML, and pretending otherwise would need a server. Use your browser’s View Source, or a fetch from your own terminal, and paste the result.
Retrieving pages on your behalf — crawling, extraction and storage into a research library — is part of the FlexContext product, where there is a backend allowed to make the request.
Questions
HTML extractor: common questions
Can I paste a URL instead of HTML?
No. Browser security prevents this page from fetching another site’s HTML, and there is no server here to do it. Paste the page source instead.
Does it convert pages to Markdown?
It returns clean readable text and structured tables rather than full Markdown. For most uses — chunking, token counting, quoting — that is the part that matters.
Is this a web scraper?
It is the extraction half of one. The fetching half, which is what site owners and rate limits care about, is deliberately not here.