A glossary of AI terms.

Short, plain definitions of the terms used around AI agents and AI operating systems, each with how it works, its limits, and how Venta uses it.

• Agentic AIAgentic AI is AI that pursues a goal with some independence: it plans, takes actions with tools, checks the results and adjusts, instead of waiting for a prompt at every step.
• AI agentAn AI agent is an AI model that works toward a goal in steps, choosing and using tools such as search, files or commands along the way, instead of giving a single answer.
• AI memoryAI memory is information an AI system keeps about you and your work between conversations, such as preferences, facts and past decisions, and brings back when it is relevant.
• AI operating systemAn AI operating system is software that sits between you and AI models and gives them what an operating system gives programs: access to files, tools, devices and memory, under rules you set, so you can hand over a task instead of only asking a question.
• Bring your own key (BYOK)Bring your own key, or BYOK, means an app sends your requests to an AI provider with an API key from your own account instead of the app's.
• Computer useComputer use is an AI acting on a real computer for you, either through the computer's own interfaces, such as commands and files, or by looking at the screen and working the mouse and keyboard.
• Context windowA context window is the amount of text, measured in tokens, that an AI model can take in at once: the instructions, the conversation so far, any documents, and room for its reply.
• Model routingModel routing is choosing, for each request, which AI model answers it, by rules such as ability, cost, speed, or which models are available at that moment.
• Multimodal AIMultimodal AI is AI that takes in or produces more than one kind of data, such as text, images, audio or video, instead of text alone.
• Retrieval-augmented generation (RAG)Retrieval-augmented generation, or RAG, is a way of answering from a body of documents: the system first finds the passages most relevant to a question, then gives them to the model to answer from.