Skip to content
Get the free beta

Atlas (.atls)

Documentation sections

An open archive format for structured web data

Atlas (.atls) is an open file format for packaging structured web analysis data alongside the raw content it was derived from. It is designed to be producer-agnostic: any crawler, scraper, or analysis tool can emit a valid Atlas archive, and any reader can consume one.

New to the format? Start with the plain-words overview: Atlas, one open file for everything a crawl finds.

An Atlas archive is self-describing: the schemas that define every field travel inside the file, so the archive can be opened and understood years later with nothing but the file itself and a ZIP and Zstandard decoder. The specification is public, so the format is not tied to any single application that creates or reads it.

What an Atlas archive contains

An .atls file is a ZIP archive. Inside: a machine-readable manifest, structured datasets stored as compressed JSON Lines, content-addressed blobs for raw page bodies, and a JSON Schema for each dataset. The manifest identifies the records and content retained by that capture; an archive is not necessarily a complete offline copy of a website. See Archive anatomy for the full layout.

Use cases

  • Technical SEO auditing and reporting
  • Captured accessibility observations
  • Before/after crawl diffs for site migrations
  • Content inventory and metadata extraction
  • Performance and response-time baselines