Lara (Local Archive, A) wants to be(come) a
- decentralized
- web search engine
- web archive
- and feed reader.
Project status: incomplete, experimental
A user should be able to run an instance of Lara on their own (or rented) hardware and connect it with other trusted remote instances.
A search query combines results from own and remote instances.
Instances only crawl and index sites handpicked and trusted by their administrators. This eases the implementation:
- Crawled sites can be assumed to not be spam or intentionally misbehaving.
- Ranking of sites is not a priority anymore since all of them are considered valuable. Thus a big webgraph and pagerank is not needed.
- Most Instances run on one single machine but instances can also scale to small clusters.
It is widely acknowledged, that the quality of Google search has declined:
- The term "enshittification"1 has been created for this phenomenon.
- lightning talk at Chaos Communication Congress 20232
The paper GOGGLES: Democracy dies in darkness, and so does the Web by the Brave Search Team discusses the issue of biases in search engines.
The Trusted News Initiative (Wikipedia!) is an alliance of Big Tech and others to combat what they consider disinformation.
See also my blog post Rebuild search with trust with further motivation and links to previous decentralized search projects, especially YaCy.
The project should also serve as an archive:
“Who controls the past controls the future: who controls the present controls the past” --- 1984, George Orwell
LaraBot is the crawler of this project. It tries to be polite. Please be patient. Open an issue in this project or contact larabot@koch.ro for any feedback. Feedback is welcome!
- Github? Git repos can be moved quickly. Github provides visibility.
- "Lara"? Provisional name. My second dog.
Footnotes
-
2023-07-28, Cory Doctorow: Microincentives and Enshittification ↩
-
2023-12-28, Martin Hamilton: Honey I federated the search engine - finding stuff online post-big tech ↩