# Glossary

> Canonical: https://www.overmindlab.ai/glossary

The terms AI teams meet when they trace, evaluate and train agents, defined plainly. Listed alphabetically.

## Terms

- [LLM observability](https://www.overmindlab.ai/glossary/llm-observability): LLM observability records every prompt, completion, tool call and retry your AI application makes as traces, so you can explain any single run. It shows what happened; judging whether the output was good is a separate job called evaluation.