What Atlas can reach, and on whose terms.
Atlas reads news from feeds and records from document collections. This page lists the collections: who publishes each one, the licence that governs it, roughly what it covers, and what it is known to miss. A gap listed here is a gap in the collection or in Atlas's reach — never evidence that a record does not exist.
15 of 16 collections were probed on 2026-09-29, and 7 have a confirmed endpoint. The rest need an API key, answered with something other than the data, or expose no machine interface at all — each row says which. A confirmed endpoint means it answered correctly once; a confirmed search is a further fact, and the collections that have one are searchable at document search. A separate index records what governments decline to release at the withheld record.
GovInfoOpenneeds an API key
U.S. Government Publishing Office · US · 1990s–present
United States federal government works, generally not subject to domestic copyright
Probed 2026-09-29: HTTP 401: the endpoint answers, but not to an unauthenticated caller
https://api.govinfo.gov/collections
- ·digitisation of older material is partial and varies by publication
- ·unknown — publication-level coverage not yet surveyed
Federal RegisterOpenendpoint confirmed
U.S. Office of the Federal Register · US · 1994–present
United States federal government works, generally not subject to domestic copyright
Probed 2026-09-29: HTTP 200, application/json; charset=utf-8, marker present
https://www.federalregister.gov/api/v1/documents.json?per_page=1
- ·issues before the digital era are available only as scanned images
National Archives CatalogOpenanswered with the wrong thing
U.S. National Archives and Records Administration · US · unknown–present
predominantly United States federal government works; some holdings carry third-party rights
Probed 2026-09-29: expected application/json, got text/html
https://catalog.archives.gov/api/v2/records/search?q=atlas&limit=1
- ·only a fraction of physical holdings are digitised
- ·unknown — digitised proportion not yet surveyed
CIA declassified records (CREST)Opena web page, not an interface
U.S. Central Intelligence Agency · US · unknown–unknown
United States federal government works, released through declassification review
Probed 2026-09-29: HTTP 200, text/html; charset=utf-8 - a page answered, but this corpus exposes no machine interface
https://www.cia.gov/readingroom/search/site/atlas
- ·contains only what declassification review has released
- ·released documents are frequently redacted in part
Federal agency FOIA reading rooms and request logsOpenneeds an API key
U.S. federal agencies, individually · US · unknown–present
United States federal government works; released records and request logs are published by each agency
Probed 2026-09-29: HTTP 403: the endpoint answers, but not to an unauthenticated caller
https://api.foia.gov/api/annual-report-data-set/
- ·each agency publishes independently, in different formats, with different completeness
- ·a released record may be withheld in part, and the withholding is the evidence
- ·unknown — no agency-by-agency survey yet
EUR-LexOpen · attribution requireda web page, not an interface
Publications Office of the European Union · EU · 1950s–present
European Union reuse policy permits reuse with acknowledgement of source; some documents carry third-party rights
Probed 2026-09-29: HTTP 202, text/html; charset=UTF-8 - a page answered, but this corpus exposes no machine interface (202 Accepted on a GET, possibly an interstitial rather than a result)
https://eur-lex.europa.eu/search.html?type=quick&text=atlas
- ·language coverage varies by document and era
- ·unknown — per-document rights exceptions not yet surveyed
UN Digital LibraryOpen · attribution requiredanswered with the wrong thing
United Nations · UN · 1945–present
reuse terms vary by document; many UN documents are freely reproducible with attribution
Probed 2026-09-29: expected application/xml, got text/html; charset=UTF-8 (HTTP 202)
https://digitallibrary.un.org/search?p=atlas&of=xm
- ·digitisation of early material is incomplete
- ·unknown — which document classes permit reproduction has not been surveyed
arXivOpen · attribution requireddid not answer
Cornell University · global · 1991–present
per-article: arXiv's own distribution licence or a Creative Commons variant chosen by the author
Probed 2026-09-29: HTTP 429: rate limited on a single request, which is itself worth knowing
https://export.arxiv.org/api/query?search_query=all:atlas&max_results=1
- ·preprints, so inclusion implies no peer review
- ·subject coverage is concentrated in physics, mathematics and computer science
- ·per-article licences differ and must be read per article
PubMed Central Open Access SubsetOpen · attribution requiredendpoint confirmed
U.S. National Library of Medicine · US · unknown–present
per-article Creative Commons or similar open licences; only the Open Access subset permits redistribution
Probed 2026-09-29: HTTP 200, text/xml; charset=UTF-8, marker present
https://eutils.ncbi.nlm.nih.gov/entrez/eutils/esearch.fcgi?db=pmc&term=atlas&retmax=1
- ·the Open Access subset is a minority of PubMed Central as a whole
- ·biomedical scope only
OpenAlexOpen · attribution requiredendpoint confirmed
OurResearch · global · unknown–present
open scholarly metadata; the metadata is openly licensed, the underlying articles are not
Probed 2026-09-29: HTTP 200, application/json, marker present
https://api.openalex.org/works?per-page=1
- ·metadata and citation graph only — it holds no document text
- ·coverage of non-English and non-indexed venues is weaker
Directory of Open Access JournalsOpen · attribution requiredendpoint confirmed
DOAJ · global · 2003–present
open metadata; indexed journals carry their own per-article licences
Probed 2026-09-29: HTTP 200, application/json, marker present
https://doaj.org/api/search/journals/atlas?pageSize=1
- ·an index of journals, not a document store
- ·inclusion depends on journals applying
Wayback MachineMeteredendpoint confirmed
Internet Archive · global · 1996–present
access is free and rate-limited; captured content remains under its original publisher's rights
Probed 2026-09-29: HTTP 200, application/json, marker present
https://archive.org/wayback/available?url=example.com
- ·a capture is an incomplete observation, never a backup
- ·crawl intervals leave gaps, and dynamic or login-gated pages are frequently missing
- ·absence of a capture is never evidence that a page did not exist
Common CrawlMeteredendpoint confirmed
Common Crawl Foundation · global · 2008–present
the crawl is freely available; crawled content remains under its original publisher's rights
Probed 2026-09-29: HTTP 200, application/json, marker present
https://index.commoncrawl.org/collinfo.json
- ·a sample of the web per crawl, not a complete index
- ·bulk retrieval carries real egress cost
CourtListener and the RECAP archiveMeteredendpoint confirmed
Free Law Project · US · unknown–present
free to read; the underlying PACER system charges per page, so completeness depends on what contributors have already purchased
Probed 2026-09-29: HTTP 200, application/json, marker present
https://www.courtlistener.com/api/rest/v4/courts/?page_size=1
- ·holds what contributors have fetched from PACER, not every filing
- ·sealed filings are absent by design, though the docket entry recording the seal is public
ACLED conflict event dataRestricteddid not answer
Armed Conflict Location & Event Data Project · global · 1997–present
licensed for academic and non-profit use on terms separate from commercial use; commercial use requires its own licence, which is an open Atlas launch blocker
Probed 2026-09-29: fetch failed
https://api.acleddata.com/acled/read?limit=1
Atlas may show the citation and metadata only — never the text.
- ·coverage depends on reported events, so under-reported conflicts are under-represented
- ·already live in the Atlas news pipeline — its commercial standing must be confirmed before any paid tier
JSTORRestrictednot checked
ITHAKA · global · unknown–present
licensed archive reached through institutional subscription; redistribution of article text is not permitted
Atlas may show the citation and metadata only — never the text.
- ·access depends on an institutional subscription Atlas does not hold
- ·citation and metadata only — Atlas may never display article text from here
Licence descriptions summarise the governing terms and are not legal advice. Each must be read at the source before Atlas relies on it.