Open knowledge bundle generator
The Open Knowledge Bundle Generator turns one public page into a short Markdown file an agent or retrieval system can read. It fetches the page, takes its title, meta description, first H1 and up to 100 links, and writes them into a bundle with the source URL, a timestamp and a clear "generated, not verified" status line.
- Reads
- one page per run
- Output
- Markdown, copy or .md
- Cost
- free, no signup
What the bundle draws on
Open any line for why it matters and how to fix it
-
Not runTitle tag—
Why it mattersThe <title> becomes the bundle's heading, so it is the first thing a reader or agent sees. If it is missing, the bundle falls back to the H1.
How to fixGive the page a specific title that names its subject, not just the brand. The Meta Tag Generator estimates whether it fits in a search result.
-
Not runMeta description—
Why it mattersThe description becomes the bundle's summary paragraph. Without it, the bundle says no description was found and the reader has no one-line overview.
How to fixWrite a factual one- or two-sentence description of what the page covers. Keep claims to what the page itself supports.
-
Not runPrimary heading (H1)—
Why it mattersThe first H1 is recorded as the page's primary heading. It should agree with the title; a mismatch suggests the page is about two things.
How to fixKeep one H1 that states the page's subject in plain words, and make sure it matches the title's meaning.
-
Not runLinked resources—
Why it mattersUp to 100 unique links become the bundle's resource list, so a reader can follow the page's references. No links means a dead-end bundle.
How to fixLink to the canonical pages that support the content, then prune navigation and low-value links from the bundle before you publish it.
Your Markdown bundle
Title, source, timestamp, status, description, primary heading and linked resources
Nothing to show yet. After a run, the full Markdown bundle appears here with copy and download buttons. It is marked Status: generated, not verified until you review it.
How to build a bundle you can stand behind
A bundle helps an agent navigate key public information without treating the whole site as one undifferentiated document. It must stay traceable to its source.
-
Step 1
Choose the right source page
Pick one focused, public page that is the authoritative record for its topic. Avoid account areas, search results, duplicate archives, private files and any URL that should not be treated as authoritative.
-
Step 2
Generate the bundle
Paste the URL and run it. The generator fetches the HTML from our server and reads the title, meta description, first H1 and up to 100 unique links.
-
Step 3
Check every generated line
Compare the title, description, heading and links with the live page. Remove links that are navigation or irrelevant, and correct anything the page itself gets wrong.
-
Step 4
Publish with the status changed
Once reviewed, change the status line yourself and publish the file where agents can find it, alongside the canonical page it came from.
-
Step 5
Regenerate when the page changes
Keep the canonical page as the primary record. Regenerate or version the bundle whenever the source changes, so an old export never becomes the authority.
What goes into the bundle
The bundle is plain Markdown in a fixed order. Every line comes from the page's HTML or from the run itself, and nothing is rewritten or summarised by AI. Here is where each part comes from:
For what the format is and who reads it, read our explainer on the Open Knowledge Bundle standard.
| Bundle section | Taken from | If it is missing |
|---|---|---|
| # Heading | The page's <title>, or the H1 if there is no title. | Falls back to "Knowledge bundle". |
| Source | The URL you entered. | Always present. |
| Generated | The time your browser built the bundle, in ISO 8601 format. | Always present. |
| Status | Fixed text: generated, not verified. | Always present. Change it only after review. |
| Description | The meta description. | Says no meta description was found. |
| Primary heading | The first H1 on the page. | Says no H1 was found. |
| Linked resources | Up to 100 unique absolute or root-relative links, one per line. | The list is empty. |
What the generator does
- Fetches one public page from MoxSEO's server
- Reads the title, meta description, first H1 and up to 100 unique links
- Builds the Markdown in your browser with source, timestamp and status
- Lets you copy it or download it as a .md file
What it does not do
- Run JavaScript, so content rendered by scripts is not seen
- Crawl a sitemap or more than one page
- Verify any claim on the page
- Read body text, images or structured data
Bundle, llms.txt or sitemap?
They do different jobs. A knowledge bundle summarises one page and its references. An llms.txt file is a proposed Markdown file at the root of a site that points language models to its most useful pages. An XML sitemap lists URLs for search engine crawlers. Many sites will want an llms.txt and a sitemap; a bundle is useful when one page carries facts an agent needs in a compact form.
The bundle format is MoxSEO's own simple layout, not a ratified standard, so no AI system is known to look for it by name. Its value is that it is short, traceable to a source URL and honest about being unverified. A bundle cannot repair unsupported claims, missing context or unclear headings on the source page; fix those first with the AI Answer Extractability Checker.
Important limitation: only include information you have the right to process and redistribute. A generated bundle does not give you permission to copy third-party content, disclose private material or publish personal information without an appropriate basis.
| File | Scope | Made for | Format |
|---|---|---|---|
| Knowledge bundle | One page and its links | Agents and retrieval systems reading one topic | Markdown |
| llms.txt | A curated list of a site's key pages | Language models deciding what to read | Markdown at /llms.txt |
| XML sitemap | Every URL you want crawled | Search engine crawlers | XML |
Review the source before bundling?
Check the source pages, metadata and permissions before turning public information into a knowledge bundle.
Questions about knowledge bundles
Short, standalone answers. Email editorial@moxseo.com if any of them is wrong.
What is an open knowledge bundle?
In this tool, it is a short Markdown file that records a page's title, source URL, generation time, verification status, meta description, primary heading and linked resources. It gives an agent or retrieval system a compact, traceable summary of one page instead of the full HTML.
What does the generator read from the page?
It reads the title tag, the meta description, the first H1 and up to 100 unique absolute or root-relative links in the HTML. It does not read body text, images or structured data, and it does not run JavaScript.
Why does the bundle say "generated, not verified"?
Because nothing in it has been checked. The generator copies what the HTML says, including any mistakes. Review every line against the live page, remove irrelevant links, and change the status yourself only after you have checked it.
Can it bundle a whole site?
No. It reads one page per run and does not crawl a sitemap. To cover several pages, run it on each one, or use the llms.txt Generator to draft a site-level list of key pages.
Is a knowledge bundle the same as llms.txt?
No. llms.txt is a proposed file at a site's root that points language models to its key pages. A knowledge bundle summarises one page and its links. The bundle layout is MoxSEO's own format, not a ratified standard.
Why is my bundle almost empty?
The generator reads the HTML your server returns and does not run JavaScript. If the title, description, heading or links are added by scripts, it will not see them, and neither will many AI fetchers. Server-render the key content and run it again.
Can I bundle someone else's page?
The tool will fetch any public URL, but you should only publish or share information you have the right to redistribute. A bundle does not give permission to copy third-party content or to publish personal information.
Run the next check
All 24 free tools- Person and Organization Schema GeneratorAdd truthful entity JSON-LD the visible page supports.
- Agentic Resource Discovery CheckerCheck whether agents can find the resources you publish.
- AI Answer Extractability CheckerMake the source page itself easy for AI to quote.
- llms.txt Generator and ValidatorDraft a site-level list of key pages for language models.
- Meta Tag GeneratorWrite the title and description the bundle is built from.
Rather we structure it
for you?
A free 30-minute call with Sakshi Kumari or Ashish Khan, the specialists who run MoxSEO. Bring the pages you want agents to read; they will tell you what to fix on the source before anything is bundled.
A call is not an engagement. We never guarantee rankings or AI citations.