CiteSeer(X)

CiteSeerX is a free, publicly accessible digital library and search engine primarily aimed at academic and scientific papers, with a strong focus on computer science and information technology. It was developed by the Pennsylvania State University College of Information Sciences and Technology, and uses autonomous citation indexing: it automatically extracts article metadata and citation links to build a searchable corpus. 

Who it serves & how

• Researchers, students and academics in computer science, information science and related fields use CiteSeerX to search for papers, track citations and explore author networks.
• It is particularly helpful when you want free access to meta-data rich collections of scholarly articles and want to examine citation relationships (e.g., who cites whom) rather than simply performing a keyword search.
• For example, a researcher could use CiteSeerX to see how their work is being cited, discover related works and even access full text (where freely available) without paywalls.

Key features & value

• Automated metadata extraction: titles, authors, abstracts and references are harvested automatically from publicly available articles. 
• Citation-linking: you can follow citations (“cited by” lists) and go through networks of papers. 
• Search includes full-text indexing (where available), which helps uncover papers based on content, not just metadata. 
• Free and open access: the service does not require subscription and makes metadata available under open licenses for non-commercial reuse. 

Considerations & limitations

• Coverage is limited compared to commercial citation databases: CiteSeerX mainly harvests papers that are publicly accessible on the web (author pages, open access repositories) and does not always include pay-walled publisher versions. 
• Because harvesting and metadata extraction are automated, errors in author names, dates or citation counts do occur. 
• While strong in computer science & IT, its coverage in other disciplines may be less comprehensive than dedicated domain databases.