What is RAG?
Retrieval-Augmented Generation
It is a method that allows artificial intelligence to produce answers by obtaining information from your current and private data.
Overview
RAG is a technique that allows artificial intelligence to respond to current or special information that is not in its own memory by accessing it from external sources. Thanks to this method, the model goes beyond the information it has memorized and creates more accurate and reliable answers based on the documents you provide.
How it works
The system first understands the question you ask and scans the relevant documents. Then, it presents the information it finds to the artificial intelligence like a note paper. Artificial intelligence reads these notes and prepares a clear answer for you with its own language ability.
Where it is used
It is used in internal company documents, technical manuals or wherever current news needs to be followed. It is especially preferred to reduce the risk of providing false information.
Commonly confused with
It is often confused with training the model itself, whereas RAG does not train the model but only makes it read an external reference.
Frequently asked questions
Does artificial intelligence learn on its own when using RAG?
No, RAG only fetches external information for the current question, it does not add anything to the permanent memory of the model.
Is internet connection required?
No, if your documents are stored on your own computer or server, they can work without the need for the internet.
Related terms
Related tools
This explanation was written in plain language for TreScout and machine-translated from the Turkish original · the Turkish version prevails. If something looks wrong or missing, write to hello@trescout.com. Read in Turkish →