BANKS is a system that enables keyword-based search and browsing of relational databases without requiring users to know the schema or write complex queries. It models tuples as nodes in a graph, connected by links induced by foreign keys and other relationships. Answers to queries are modeled as rooted trees connecting tuples that match individual keywords. Answers are ranked using proximity and prestige, similar to techniques used in Web search. The system provides a rich interface for browsing data and automatically generates hyperlinks based on foreign keys and other links. BANKS is implemented in Java using servlets and JDBC and can be run on any schema without programming. It is accessible via the URL http://www.cse.iitb.ac.in/banks/. The system models the database as a graph with nodes representing tuples and edges representing relationships. Node weights and edge weights are used to determine relevance. Node weights are based on the prestige of the node, while edge weights reflect the strength of the relationship between tuples. The system uses a heuristic algorithm to find and rank query results. It supports keyword searches and browsing, allowing users to extract information by typing keywords, following hyperlinks, and interacting with controls on the displayed results. BANKS provides a framework for answering keyword queries, with a model that incorporates proximity and prestige. It allows users to search for information in a simple manner without any knowledge of the schema or query languages. The system has been evaluated using academic and bibliographic databases, and it has been found to return intuitive and useful answers. The system is practical for moderately large databases and reduces the effort involved in publishing relational data on the Web and making it searchable. Examples of data that can be published using BANKS include organizational data, bibliographic data, and product catalogs. The system also supports metadata queries and has been shown to be effective in handling such queries. Future work includes extending the system to handle XML data and improving query evaluation techniques.BANKS is a system that enables keyword-based search and browsing of relational databases without requiring users to know the schema or write complex queries. It models tuples as nodes in a graph, connected by links induced by foreign keys and other relationships. Answers to queries are modeled as rooted trees connecting tuples that match individual keywords. Answers are ranked using proximity and prestige, similar to techniques used in Web search. The system provides a rich interface for browsing data and automatically generates hyperlinks based on foreign keys and other links. BANKS is implemented in Java using servlets and JDBC and can be run on any schema without programming. It is accessible via the URL http://www.cse.iitb.ac.in/banks/. The system models the database as a graph with nodes representing tuples and edges representing relationships. Node weights and edge weights are used to determine relevance. Node weights are based on the prestige of the node, while edge weights reflect the strength of the relationship between tuples. The system uses a heuristic algorithm to find and rank query results. It supports keyword searches and browsing, allowing users to extract information by typing keywords, following hyperlinks, and interacting with controls on the displayed results. BANKS provides a framework for answering keyword queries, with a model that incorporates proximity and prestige. It allows users to search for information in a simple manner without any knowledge of the schema or query languages. The system has been evaluated using academic and bibliographic databases, and it has been found to return intuitive and useful answers. The system is practical for moderately large databases and reduces the effort involved in publishing relational data on the Web and making it searchable. Examples of data that can be published using BANKS include organizational data, bibliographic data, and product catalogs. The system also supports metadata queries and has been shown to be effective in handling such queries. Future work includes extending the system to handle XML data and improving query evaluation techniques.