About this dataset
Tracks the npm publish firehose: npm-newest carries the newest package updates (release events for dependency monitoring), npm-unscoped highlights unscoped package names (rare, discovery-relevant — most modern packages are @scoped), and npm-summary carries corpus stats. The npm updates feed is event-driven: each item is a package publish with its event timestamp and current README excerpt. Rows are intentionally lean: package name, canonical npmjs.com URL, publish time, publisher, README excerpt. Source: registry.npmjs.org (keyless official feed, updates continuously). Pipeline is idempotent with content-fingerprint dedupe (package + URL + publish event time); corpus freshness is anchored on-chain each hourly run so buyers can verify data age without trusting us.