One open file for captured website evidence.
Atlas is the archive Cartographer writes and Continuum reads: collected pages, links, responses and available screenshots, kept together in a single .atls file that belongs to you.
Open specification. Reader SDK on npm.
One file per crawl.
An Atlas retains the evidence collected within a crawl’s scope. Check its manifest, datasets and diagnostics to understand coverage, skipped pages and unavailable measurements. It is a record of that capture, not a complete copy of the live website.
Continuum uses it to find issues, follow a finding back to the page, and compare one crawl with the next. Other readers can inspect the stored data. Native comparison eligibility and analysis depend on the application and compatible capture evidence.
- Captured pages, as readable text and as the source it was built from.
- Collected links, with where on the page it sat and where it went.
- Recorded responses, with its status, headers, and timing.
- Screenshots, structured data, and accessibility checks, where the collection mode and run produced them.
- The schemas that describe all of it, inside the same file.
Yours to keep.
An open specification.
The format is published under an open license (CC BY 4.0). Anyone can write an Atlas, and anyone can read one, with or without our software.
Readable without Continuum.
The reader SDK is open source (Apache-2.0) and on npm. Underneath, an Atlas is a ZIP of documented, compressed tables, so there is nothing to be locked into.
Still readable years from now.
The schemas that define every field travel inside the archive. The stored schemas describe the recorded datasets. Use a compatible reader or the documented decompression and table-reading tools.
Written by Cartographer. Read by Continuum.
The file in the middle is yours either way.
Inspect an example, then go deeper.
Explore fictional reference captures online and download the exact files. Each release identifies its producer, coverage and compatibility limits.
Anatomy of an archive
How an Atlas is put together: the manifest, the datasets, the blobs, and the schemas that travel with them.
What each dataset holds
Pages, links, responses, assets, structured data, accessibility, and more, field by field.
Atlas SDK
Open and read any Atlas from TypeScript or JavaScript. Install from npm and start with a few lines.
Follow Continuum releases.
Sign up for product and availability news. The free beta is one email away on the beta page.