AI Glossary · 22 terms · last reviewed May 2026

Plain-English AI definitions, written by humans.

Every entry is hand-written, tied to the real tools that use the concept, and updated at least every six months. No auto-generated dictionary entries.

22Terms defined

100%Hand-written, no AI slop

LinkedTo the tools using each

May 2026Last full review

AI Agents

A program that takes goals and figures out the steps to reach them, calling external tools to act in the worlde.

API

Application Programming Interface

The way one piece of software talks to another.

Chain of Thought

A prompting technique where the model reasons out loud, step by step, before giving an answer.

Context Window

How much text a model can read at once.

Embeddings

Numeric fingerprints of text or images that let computers measure semantic similarity.

Few-shot Learning

Showing a model two to five examples in the prompt so it follows the pattern, without any retraining.

Fine-tuning

Training a pre-built model further on your own data, to make it better at a specific task or voice.

Hallucination

When a model confidently states something false.

Inference

Running a trained model to get a prediction or response.

Latency

Time-to-first-token: the delay between sending a request and seeing the first word.

LLM

Large Language Model

A neural network trained on huge amounts of text, capable of generating, summarising, and reasoning about language.

MCP

Model Context Protocol

An open standard that lets AI models connect to external tools, data sources and services in a consistent way.

Multimodal

A model that handles more than text, like image, audio or video.

Prompt Engineering

Writing the instructions to a model in a way that consistently produces useful output.

RAG

Retrieval-Augmented Generation

A method that fetches relevant documents at query time and feeds them to the model, so answers are grounded in real sources.

RLHF

Reinforcement Learning from Human Feedback

A training method where humans rank model outputs to teach the model what good looks like.

Semantic Search

Search that matches on meaning instead of exact words.

System Prompt

The hidden instruction that sets a model behaviour and constraints before a user message arrives.

Tokens

The chunks (usually word-pieces) that models read and produce.

Transformer

The neural network architecture behind almost every modern LLM.

Vector Database

A database optimized to store and search embeddings.

Zero-shot Learning

A model performing a task it was never explicitly trained on, just from instructions.

How we write these definitions

Hand-written, every entry
A real person writes each definition, drafts the examples and links to the actual tools that use the concept. No auto-generated content.
Reviewed every 6 months
AI moves fast. We re-check every entry at least every six months and rewrite ones that have drifted.
Tied to real tools
Every entry ends with hand-reviewed tools that use the concept, so you can go from definition to working software in one click.

Spot a missing term?

If a concept matters in real AI work and we have not defined it, tell us. We will write the entry and credit you on the page.

Suggest a term

Want to be everywhere, not just here?

Get listed on 270+ directories, by hand.

Our sister service, FreeAIdirectories, submits your tool to 270+ AI directories by hand.

Submit to 270+ directories

270+

Directories we cover

100%

Hand-submitted, not bots

Pay-to-rank slots, ever

Free

To list here, with dofollow