Every week AI pulls a new acronym out of its sleeve. Agentic this, RAGRAG (Retrieval Augmented Generation)Combine real data search with text generation that, vibe codingVibe CodingProgram by describing what you want and letting the AI write the code, MCPModel Context Protocol (MCP)Standard protocol for models to connect with external tools… and everyone nodding on LinkedIn as if they had it all figured out.
At NAiOS Labs we got tired of searching for the same definition across fifteen different tabs, between papers, X threads, and 40-minute videos to explain what fits in two sentences. So we did the logical thing: we built our own wiki. And today we are opening it to the public.
Introducing the NAiOS Wiki
An open glossary of AI terminology: concepts, techniques, architectures, and trends, explained clearly, to the point, and without the hype. Each term comes with a one-sentence definition and, below it, an in-depth explanation for when you want to dive deeper. No fluff, no sales pitches.
Nine areas, from the technical to the cultural
We have organized it into our own categories so you can find what you are looking for… or get lost happily jumping from one idea to another:
🤖 Agents and Autonomy🧠 Reasoning and Prompting🧩 Architecture and Models🗄️ Training and Data✨ Content Generation🛡️ Safety and Ethics🏛️ Governance and Strategy🖥️ Infrastructure and Deployment📈 Trends and Culture
Designed to actually be read
Definitions that are understood the first time, without needing another dictionary to understand the dictionary.
Related terms in each entry to pull the thread and connect concepts.
Search engine, A-Z index, and filters by category to get straight to the point.
And a touch we love: when a term appears in blog posts, you will see it underlined — hover your mouse over it and you will get its definition without leaving the article.
Generalist and honest. The NAiOS Wiki is an educational and independent glossary, not a platform brochure. Some terms are implemented by us; others —including risks and bad practices like prompt injectionPrompt InjectionAttacks that manipulate AI by inserting malicious instructions or deepfakes— are there just for you to understand them, not because we sell them.
A living project: more every week
AI doesn't stop, so neither do we: we add new terms every week, as models, techniques, and buzzwords appear. What you see today is just the starting point; the wiki is designed to grow with you.
Missing a term?
This is where you come in. If there is an acronym or a concept that isn't here yet, send us your suggestion via the contact form and our team will review it for inclusion. The best wikis aren't written by a company: they are built by a community.
Stop pretending you understand that acronym; new ones appear every week and only the most relevant ones go down in history
Explore the wiki now and come back when the next trend catches you off guard.
