Skip to content
WorldFactbooks

Where to Download the CIA World Factbook Archive

Published

The CIA shut down The World Factbook on February 4, 2026, with the final content update published January 22, 2026. No official government archive exists: cia.gov redirects every old Factbook URL to a farewell story, and neither GPO nor NARA took over hosting. The data survives anyway, in several independent archives, and because it is a US government work it is public domain and free to download, reuse, and republish.

This guide covers the four main download routes, what each contains, and which format fits which job. Our own site builds on some of these sources; where another archive is the better tool for your task, we say so.

factbook.json: the final edition as clean JSON

The factbook/factbook.json project on GitHub converted the Factbook’s country pages to JSON weekly from 2021 until the source disappeared. Its data is frozen at the final CIA update of January 21-22, 2026, which makes it the canonical machine-readable copy of the last edition.

What you get: 261 country JSON files, about 9.4 MB of raw data, organized into 12 regional directories (africa/, europe/, south-asia/, and so on) plus world/ and meta/. Each file contains the full profile: 13 top-level categories (Geography, People and Society, Government, Economy, Energy, and others), with each field’s value as a display string, for example "1,419,316,933 (2025 est.)" for India’s population.

License: CC0 1.0 on top of the underlying public-domain status. No restrictions, no attribution required.

How to pull it:

# Full repo (about 67 MB with history)
git clone https://github.com/factbook/factbook.json.git

# Or just the data as a tarball (about 11 MB extracted)
curl -L -o factbook.tar.gz \
  https://codeload.github.com/factbook/factbook.json/tar.gz/refs/heads/master

# Or a single country, raw
curl https://raw.githubusercontent.com/factbook/factbook.json/master/south-asia/in.json

One trap: files are named by GEC codes (the CIA’s legacy country codes), not ISO codes. ch.json is China, not Switzerland. gm.json is Germany, not Gambia. ja.json is Japan, and kr.json is Kiribati, not South Korea. Never guess a filename from ISO habits; check the repo’s directory listing or a GEC cross-reference first.

Good for: anyone who wants the final edition’s text in a simple, predictable structure. Less good for numeric analysis, because every value is a formatted string (“$14.244 trillion”) that you would have to parse yourself. For that, use the sibling repo below.

cache.factbook.json: the same data with parsed numbers

The lesser-known sibling, factbook/cache.factbook.json, stores the original JSON that cia.gov itself served (the final Factbook site was an application backed by structured JSON). Same 261 entities, same January 2026 vintage, same CC0 license, but the files carry much more machine-usable detail:

  • A parsed numeric value for quantitative fields: India’s GDP (PPP) appears as "14244000000000" alongside the display string “$14.244 trillion (2024 est.)”, with prefix and suffix fields for units.
  • estimated flags and per-field date information.
  • Per-country published and updated timestamps, so you know exactly when the CIA last touched each page.
  • Stable numeric field_id values that are consistent across countries, plus a comparative flag marking which fields the CIA considered comparable between countries.
  • media arrays referencing flags, maps, and photos.

Files run about three times larger (India is 192 KB versus 57 KB in factbook.json). The repo is around 186 MB with git history, so grab the tarball instead of cloning:

curl -L -o cache-factbook.tar.gz \
  https://codeload.github.com/factbook/cache.factbook.json/tar.gz/refs/heads/master

Good for: building anything computational: rankings, comparisons, charts, databases. If you plan to do math with Factbook numbers, start here and skip the string parsing entirely. Our own rankings and comparison pages are built on this class of parsed data.

The Wayback Machine: pages as they looked, plus yearly zips

The Internet Archive holds roughly 29,000 snapshots of Factbook pages going back to January 2017, with sparser coverage reaching into the 1990s. This is the only source that preserves the pages as pages: layout, maps, photos, and the exact wording on a specific date.

Single pages: prepend a Wayback prefix to any old Factbook URL. For the final-era India page:

https://web.archive.org/web/2026/https://www.cia.gov/the-world-factbook/countries/india/

For pre-2021 pages, use the old URL form, remembering the GEC codes: https://web.archive.org/web/2020/https://www.cia.gov/library/publications/the-world-factbook/geos/in.html.

Yearly zips: the CIA published downloadable full-site archives for years, and the Wayback Machine captured them. Two URL patterns to search for:

  • 2007-2021 era: cia.gov/library/publications/download/download-{YYYY}/factbook.zip
  • 2021-2026 era: cia.gov/the-world-factbook/about/archives/download/factbook-{YYYY}.zip

A recovered copy of the 2020 zip also lives in the simonw/cia-world-factbook-2020 repo, browseable as a full mirror of the 2020 site, including the useful Appendix D country-code cross-reference table.

Good for: citation repair (our /fix-citation/ tool finds the right snapshot and formats the citation for you), verifying what a page said on a given date, and grabbing whole historical editions as HTML. Not good for data analysis: you would be scraping HTML that changed format several times over the years.

worldfactbookarchive.org: 36 editions, one database

For historical depth, worldfactbookarchive.org is the most complete structured archive: 36 editions covering 1990-2025, 281 entities, and over a million parsed fields normalized into a single database. It is a non-commercial preservation project, and its bulk downloads are excellent:

  • On-site exports: CSV, Excel, and PDF, scoped from a single country-year up to the complete flat dataset.
  • Its GitHub repository (CC0 1.0) offers per-year SQL dumps (country_fields_1990.sql.gz through 2025) and a single SQLite database, factbook.db, around 662 MB with a full-text index and ISO country codes.
  • A raw-sources release (about 3 GB) preserves the original source files each edition was parsed from, with checksums.

Good for: time-series and historical research: how a country’s data changed across three decades, cross-edition comparisons, or offline analysis of the whole corpus in SQLite. If your question is “what did the 1995 Factbook say,” this is the right archive, and we link to it rather than duplicating it. What it deliberately does not offer is current data: the project’s charter is preservation, so its series end at the 2025 edition. For post-shutdown figures, use live-refreshed sources such as our country profiles (methodology here).

Which download for which job

You wantUse
Final edition, simple JSONfactbook.json
Final edition, parsed numbers for analysiscache.factbook.json
A page exactly as it appeared on a dateWayback Machine
A full historical edition as HTMLWayback yearly zips, or simonw’s 2020 mirror
1990-2025 data in CSV or SQLiteworldfactbookarchive.org
Current, refreshed figureslive sources, including this site

The World Factbook is a work of the United States government and therefore not subject to copyright under 17 U.S.C. 105. The CIA’s own (now archived) copyright page stated that Factbook data, including photos and maps, is in the public domain and may be copied freely without permission. Attribution is not legally required, though the CIA requested a courtesy citation, and the JSON repos above add an explicit CC0 dedication covering their conversion work.

The one restriction: the CIA seal. Reproducing the seal without permission is prohibited under the CIA Act of 1949, so strip it from anything you republish, and avoid implying CIA affiliation or endorsement. The data is yours to use; the insignia is not.

More guides