Back to ApunaAI agent a program that can take actions on your behalf, like sending an email or searching the web, not just answer questions Model the trained system that generates the text; what most people mean when they say "the AI." Token the small chunks a model reads and writes; roughly three quarters of a word on average, which is why AI bills are counted in tokens Prompt the instruction or question you give the model; better prompts get better results, every time API a connection point that lets two software systems talk to each other LLM Large Language Model; just another name for the kind of AI that reads and writes text Human in the loop a design choice where a person reviews or approves what the AI proposes before it acts Cloud vs on your own computer running AI on a provider's servers means lower setup cost but your data leaves your building; running it on your own hardware keeps data local but costs more upfront AI integration connecting an AI model into an existing task so it handles the mechanical reading or drafting, while a person keeps the judgement and the final decision Automation using a rule, script, or scheduled job to do repetitive work without AI; older and simpler than AI, and often the cheaper fix Process analysis making the knowledge in someone's head (which supplier tends to slip, which machine needs an extra check) visible in a dashboard or alert, so it survives even if that person is out sick or moves on RAG Retrieval-Augmented Generation; a technique that gives an AI model access to a specific set of documents at query time, so it can answer questions about your own data rather than only what it was trained on Fine-tuning adjusting an existing model on new examples so it responds in a particular style or domain; rarely necessary and often over-sold as a solution Hallucination when a model produces a confident-sounding answer that is simply wrong; the main reason a human must review AI output in any high-stakes context Guardrail a check or constraint that prevents the model from producing certain outputs On-premises (on-prem) software that runs on hardware inside your own building, under your control Open-source software whose source code is publicly available; Apache-2.0 is one common licence that allows commercial use without restriction Inference the act of running a model to generate an output; what happens when you send a message and the AI responds Context window the maximum amount of text a model can read and hold in attention at once; once you exceed it, earlier content falls out Embedding a mathematical representation of text that lets a system find semantically similar passages without exact keyword matching Repository (repo) the folder, tracked by version control, that holds a project's code and its full history of changes Commit a saved snapshot of changes to the code, with a short message explaining what changed and why Branch a parallel copy of the code where someone can make changes without affecting the main version until they are ready Pull request (PR) a proposed set of changes, opened for review before it is merged into the main codebase; the core unit of how most software teams actually ship Frontend / backend the frontend is what a user sees and clicks in a browser or app; the backend is the server-side logic and data a user never sees directly Endpoint a specific URL a program can send a request to, usually part of an API CI/CD Continuous Integration / Continuous Deployment; automated checks and steps that test and ship code changes without a person manually repeating the same steps each time Deploy to make a new version of the software live, so real users are running it Environment variable a setting (like a password or a server address) kept outside the code itself, so the same code can run differently in different places without being rewritten Container a packaged bundle of code and everything it needs to run, so it behaves the same on any machine that can run containers
Glossary
The words, explained.
AI and coding jargon, in plain language — written so a curious ten-year-old could follow it and a busy manager could use it. No sales pitch here, just the words I use on this site, defined once so you don't have to guess.