web-page-extract

Fetch a web page and return cleaned page content, metadata, and normalized links.

Why install it

Most agent systems need a reliable first-step web ingestion primitive. This tool gives you one packageable contract for page fetch + cleaning instead of rebuilding the same fetch/parse glue inside every agent.

Inputs

• url: page URL to fetch
• format: markdown or text
• include_links: include extracted links
• timeout_ms: request timeout
• max_chars: output content cap

Outputs

• url: original requested URL
• final_url: final fetched URL after redirects
• title
• canonical_url
• published_at
• byline
• excerpt
• format
• content
• links
• metadata

Local development

node --test

To build the packaged entrypoint:

npm run build

Example invocation

printf '%s' '{"url":"https://example.com","format":"markdown"}' | node dist/index.js