We have launched two new versions of Arquivo.pt: PageSearchSolr, on 14 October 2026, and Gaia, on 29 October 2026. The main novelty is a new page search: you can now find twice as many pages when you search by words. Many preserved pages that until now could only be found by their address (URL) now also appear in a text search.
There are also new ways to search, new features in the API, and more protection for the services and for the people who use them.
Twice as many pages in text search
- More pages. 5.4 billion pages and documents are now available in word search. The size of the indexed collections grew from 658 TB to 1,479 TB (+125%), and the number of indexed collections from 106 to 194 (+83%).
- New technology. Search was rebuilt on Apache Solr, replacing NutchWax. It was an effort of several years.
New ways to search
Timeline: when was a topic important?
Until now, a search returned a list of results, but did not show in which years the topic was most present on the Web. Now, bars per year above the date selector show the search's impact over time. The impact is adjusted to the size of the archive in each year, so that the years can be compared fairly.
Try it: search for "troika".
Search suggestions
If you misspell a word, Arquivo.pt suggests the correct spelling. For example: Lizboa → Lisboa, or Amalia Rodrigue → Amalia Rodrigues. Suggestions work in page and image search, and are generated by Arquivo.pt itself, from its own data, without using external services.
New criteria in Advanced Search
- Language. A new filter to search only pages in one language (for example pt, en or es), with a minimum confidence in language detection: High (default), Medium or Low. You can also use this filter directly in the search box, by typing for example
language:pt. - Collection. Search only one or several collections (for example EAWP6,EAWP8), in pages and images. This option already existed, but is now available in Advanced Search.
- Year balance (experimental). The archive holds less content from before 2015 than from after 2020, so results come mostly from the most recent years. This new option lifts the pages from the years with less content in the results. You can choose Off (default), 0.25, 0.5, 0.75 or 1. We will still evaluate its performance and the quality of the results.
- More results in image search. You can ask for up to 200 results.
For researchers and developers
The page search API supports the same options as Advanced Search (language, minimum language confidence, collection, year balance and search suggestions), and also lets you:
- get the distribution of the results over time (the
timelineparameter): the number of pages per year and the normalised impact; - choose the length of the snippets.
The API documentation has been updated and is available at arquivo.pt/api-docs. The full-text page search API (/textsearch) and the URL search API (CDX, /wayback/cdx) now have an interactive environment (Swagger) for trying out requests directly in the browser.
Protection and privacy
ArchivePageNow available again
ArchivePageNow lets you preserve a web page immediately for future consultation. As a precaution, the service was temporarily suspended after we identified its misuse by third-party automated artificial intelligence agents. All other Arquivo.pt services remained available throughout.
During this period we carried out a security audit and strengthened the service's protection with:
- an anti-bot check;
- usage limits;
- mandatory acceptance of the data protection terms.
ArchivePageNow is available again. Please remember that it is intended for educational, scientific and research purposes, in accordance with the terms and conditions.
CitationSaver has the same protections: an anti-bot check and mandatory explicit consent.
Cookie notice on archived pages
Archived pages may contain original code that uses cookies. When you view an archived page, you now see a notice about this, and the Cookie Policy has a new section that explains it.
Improvements shared with the community
Arquivo.pt uses pywb to display archived pages, open-source software used by web archives all over the world. We contributed improvements that are available to the whole community:
- Performance: faster viewing of pages with many archived versions;
- Quality: pages that redirect with a Refresh header no longer jump to the live Web, and pages compressed with Brotli are archived and displayed correctly;
- Security: more control over the addresses ArchivePageNow can visit.
Other improvements
- A new "Arquivo.pt in numbers" dashboard, with up-to-date statistics on what can be searched: pages, addresses and images.
- A new sobre.arquivo.pt. This site has been renewed: it is now a static site, faster, simpler to maintain and easier to preserve.
Arquivo.pt as a contingency plan
In October 2026, following a security vulnerability disclosed in PHP, several FCCN and FCT websites were temporarily redirected to their versions preserved on Arquivo.pt while they were being fixed. Arquivo.pt thus also served as a contingency plan to keep information accessible.
What comes next
We will keep making more collections available in text search and improving the ranking of results.
The technical details of each version are on GitHub: PageSearchSolr and Gaia.
Your suggestions are welcome!