Retrieval Augmented Generation (RAG) Software Glossary: Key Terms & Concepts

A webhook (also known as a web callback or HTTP push API) allows a programme to send real-time data to other apps. A webhook sends data to other apps as it happens, so you can get it right...read more

API integration is a feature that allows different software systems, platforms, or applications to seamlessly communicate with each other. It enables the exchange of data, functionality, a...read more

Data extraction is a crucial feature of any software that is designed to handle large amounts of data. It is the process of retrieving relevant information from a database or other sources...read more

Batch processing in background check software refers to the capability to handle multiple background screening requests simultaneously, streamlining large-scale verification processes. Thi...read more

Caps how many requests a given client, IP address or API key may make within a time window, and rejects or delays anything beyond it. It blunts brute-force attempts and scraping, and prote...read more

Keyword search matches the literal terms in a query, while vector search matches on semantic similarity of embeddings. Hybrid search runs both and merges the results into a single ranking,...read more

A retrieval method that compares numerical embeddings of content rather than matching keywords, so results are ranked by semantic similarity to the query. Text, images or audio are convert...read more

Self-hosting means the software runs inside infrastructure you control rather than the vendor, usually from an open-source build, a container image or a licensed enterprise package. For a...read more

Chunking splits a parsed document into the pieces that get indexed and retrieved, which sets the size of the context a model finally sees. Too small and a passage loses the sentence that ...read more

An embedding model turns a passage into a vector so that text with a similar meaning lands near it, which is what makes retrieval by meaning possible at all. The practical differences are...read more

A web search API gives an application live results from the open web through an endpoint, instead of a person typing into a browser. The things that separate them are freshness, whether p...read more

Grounded citations attach the retrieved passage to the generated answer, so a reader can see which source each claim came from and open it. Without them a retrieval pipeline cannot be deb...read more

Document parsing reads a PDF, scan, slide deck or spreadsheet and returns ordered text with its structure intact: headings, tables, figures and reading order preserved rather than flattene...read more

Reranking takes the candidate passages a first search returned and scores them again with a model that reads the query and the passage together, so the handful sent to the generator are th...read more