Restad (Relational Storage for Tagged Documents) is an indexing-querying tool for tagged documents using a relational database for storage and querying. It actually uses PostgreSQL because it's "the world's most advanced open source database". Restad has been developped by the Computer Science lab of the university of Avignon.
Restad provides tools to index the documents in the database and SQL functions to query the documents for full-text search on semi-structured data. First you will need to install the tools and set up the database, see Install guide.
For now, indexing tools only works on local files, that means you need to run the command on the system hosting the files. However the database can be a distant server. See How to index.
Restad does not provide any interface to query the database because there a lot of good tools to connect to and query a PostgreSQL database. The work is done by the database server, so Restad is how efficient as your PostgreSQL server is, and will always make the most of PostgreSQL updates. See details about database schema and functions in How to query.
Restad is open-source and anyone can contribute or fork and edit code to fit to special needs. See Developer guide.