Kumbukum Docs

Appearance

URLs ​

Save links to articles, documentation, tools, or any webpage. Kumbukum automatically extracts the page content so you can search your bookmarks by what they say — not just their address.

What You Can Do ​

Save a URL ​

Click New URL or paste a link. Kumbukum fetches the page and extracts:

  • Title and description from meta tags
  • Open Graph image for a visual preview
  • Full text content for search indexing

You can override the extracted title or description with your own if you prefer.

URLs are searchable across four fields: title, description, full text content, and vector embedding. This gives URLs the broadest search surface of all item types — you can find a saved link by anything that appeared on the page.

Scheduled Crawling ​

Enable crawl on any URL to have Kumbukum periodically re-fetch and re-index the page content. This keeps your saved links up to date as web pages change over time. Crawling runs automatically on a daily schedule.

For sites you want to index more deeply, full-site crawling follows internal links to capture multiple pages under the same domain.

Project Assignment ​

Every URL belongs to a project. Filter by project in the sidebar to work with a focused set of bookmarks.

Connect a URL to related notes, memories, or other URLs. Open a URL, scroll to the Links section, and search for items to connect. Links appear as removable badges and show up in the Knowledge Graph.

Batch Operations ​

Select multiple URLs with checkboxes to move them between projects or trash them in bulk.

Browser Extension ​

Use the Browser Extension to save the current page to Kumbukum without leaving your browser. Available for Chrome and Firefox.

How It Works ​

When you save a URL, the content extractor (Cheerio-based) fetches the page and pulls out structured data. The extracted content is stored in MongoDB and indexed in Typesense with a vector embedding. If extraction fails (e.g., the page blocks requests), the URL is still saved with whatever metadata you provided.

Crawl-enabled URLs are processed by a Playwright-based crawler that can navigate JavaScript-heavy pages and follow links across the site.