MosaicRAG Search Engine Framework

MosaicRAG is a framework that enables developers of search applications to build their own Retrieval-Augmented Generation (RAG) search engine. MosaicRAG uses search results coming from Mosaic or Chroma and allows to define a sequence of RAG processes (e.g. summarisation, re-ranking, chat) on retrieved search results. Such sequences can be stored as search engines for further use. Together with Mosaic it can be used as an out-of-the-box search application or as an underlying technology to create more complex applications based on the index data retrieved from the Open Web Index (OWI) of the OpenWebSearch.EU project.

Mosaic is a generic search application that makes index slices searchable that have been retrieved from the Open Web Index of the OpenWebSearch.eu project. Index slices are small or medium sized indices containing web documents related to a certain topic or for a particular purpose. These slices contain various metadata, such as geo-coordinates, topics, and genres of the web documents. More information can be found on https://doi.org/10.5281/zenodo.13790237

The software package (two zip files) contains all source files, as well as build and start scripts. A readme file describes the package, installation, and usage in detail.