Embedding

An embedding is a list of numbers representing a piece of text's meaning, positioned so that texts with similar meanings sit close together. It lets software find related content even when the wording shares no words.

Embeddings are what make semantic search possible: "how do I get my money back" and "refund policy" land near each other despite having no words in common. Convert both to embeddings, measure the distance, and the match surfaces.

The limitation is the mirror image of the strength. Because embeddings capture meaning rather than characters, they are unreliable for exact matching — model numbers, SKUs, error codes and names. Production search almost always runs keyword and embedding search together and merges the results.

Why it matters

If a search or chatbot keeps missing the one document you know exists, embeddings-only retrieval is usually the reason, and adding keyword search alongside is usually the fix.

Related terms

RAG (Retrieval-Augmented Generation)RAG is the technique of searching your own documents for passages relevant to a question, then giving those passages to a language model along with th…Vector DatabaseA vector database stores embeddings and finds the closest matches to a query quickly. It's the storage layer behind semantic search and most RAG syste…Large Language Model (LLM)A large language model is a system trained on very large amounts of text to predict likely continuations. That single ability underlies writing, summa…

More in AI Automation & Agents

AI AgentWorkflow AutomationLarge Language Model (LLM)RAG (Retrieval-Augmented Generation)Vector DatabaseHallucinationHuman in the LoopGuardrailsTool Use (Function Calling)Model Context Protocol (MCP)PromptTokenEvalsIdempotencyWebhookAPI IntegrationAI Voice Agent

Need this built rather than explained?

We publish every price we charge, and you get a quote in writing before anything starts.

See every price
All terms