Agent retrieval
dictionary.
22 compact definitions for the vocabulary behind RAG, web grounding, ranking, indexing, and agent search.
Newest glossary terms by publish or update time. This is a display group, not a CMS category.
Search APIs, query handling, and result extraction.
Web retrieval API
An API that takes a query and returns ranked web documents in a format an LLM can consume.
Hybrid search
Combining keyword-based (BM25) and vector-based (embedding) retrieval into one ranked list.
Embeddings
Dense numeric vectors that represent text in a high-dimensional space where similar texts are close.
SERP (Search Engine Results Page)
The structured list of results a traditional search engine returns to a query.
Long-tail retrieval
Returning relevant results for rare, specific, or obscure queries - not just the popular ones.
Grounding patterns that connect retrieved context to generation.
RAG (Retrieval-Augmented Generation)
A pattern where an LLM is given relevant retrieved context at inference time rather than relying solely on its training data.
Web grounding
Forcing an LLM's answer to cite fresh, retrievable web sources rather than hallucinating from training data.
Chunking
Splitting long documents into smaller passages before indexing so each piece fits in retrieval and LLM context.
Citation
A link or reference back to the source from which an LLM-generated claim was derived.
Web Grounding for LLMs
Web Grounding is the process of connecting an LLM to the live internet.
Tool-calling systems and agent runtime vocabulary.
Context window
The maximum amount of text (in tokens) an LLM can attend to at once.
AI agent
An LLM-driven system that takes actions in the world, often by calling tools and APIs in a loop.
Tool use
An LLM's ability to invoke external functions / APIs as part of generating its response.
MCP (Model Context Protocol)
Anthropic-proposed open protocol for connecting LLMs to external tools and data sources.
Agentic Search
An AI-driven search method where an autonomous agent plans, executes, and refines a multi-step search process to answer complex queries, rather than performing a single, static lookup.
Second-stage relevance and result ordering concepts.
Storage and indexing primitives for retrieval systems.
Benchmarks and measurement terms for retrieval quality.
Lost in the middle
The empirical finding that LLMs pay less attention to content placed in the middle of a long context window.
FRAMES benchmark
A retrieval evaluation benchmark testing multi-hop, freshness-sensitive web QA.
Seal-0 benchmark
A long-tail retrieval benchmark designed to expose shallow web indexes.