Quick start

Once kumo is on your PATH, crawl a host into the data tree:

kumo scrape example.com --max-pages 20

Each page is written as pages/<host>/<path>.md under your data directory: a JSON front-matter block with every structured field, followed by the page content as Markdown. Read back what you crawled, offline:

kumo pages example.com -o table

Before a full crawl, look at the frontier without fetching anything:

kumo scrape example.com --dry-run -o url     # the URLs a crawl would start from
kumo sitemap example.com                     # the same, from robots.txt and sitemaps

Work with a single page when you do not want a whole crawl:

kumo page https://example.com/ -o json | jq .title
kumo links https://example.com/              # its outbound links, as URIs

Add -o jsonl to stream records into the rest of your tools, and --limit to cap any command. See the CLI reference for the full surface.