Glossary

This glossary defines terms used in the NeuronDB documentation.

Terms

• Extension: A Postgres extension (here: NeuronDB) that adds types, functions, operators, and index methods.
• Vector: Fixed-length numeric array representing an embedding.
• Embedding: A vector representation of text/image/audio produced by a model.
• Hybrid search: Retrieval combining dense vector similarity with sparse/lexical signals and filters.
• Reranking: Second-stage scoring of candidate results to improve relevance.
• MCP: Model Context Protocol. Tool interface exposed to LLM clients.
• Agent: A service/workflow layer that orchestrates tasks using DB + tools.
• Multi-tenant: Patterns that isolate customer/tenant data within one logical system.