AI Glossary
Overview of AI and automation terms.
-
An interface that allows two applications to communicate and exchange data according to precisely defined rules.
Learn more -
A secret identifier that authorizes your application to call an external API and allows the provider to track usage.
Learn more -
Advanced AI implementations that independently decide on steps to achieve a set goal.
Learn more -
A systematic skew of an AI model originating in training data that leads to unfair or incorrect outputs.
Learn more -
A security mechanism in browsers that monitors where requests to your server can come from.
Learn more -
A prompting technique that makes an LLM reason step by step, improving accuracy on complex tasks.
Learn more -
A conversational application that replies to users in real time. Modern variants use LLMs and RAG.
Learn more -
The process of dividing long text into smaller, logical units (chunks) for efficient processing in RAG systems.
Learn more -
The maximum amount of text (in tokens) that an LLM model sees and processes at once in a single request.
Learn more -
A specialized type of AI that converts text into numerical vectors for efficient semantic search.
Learn more -
A specific URL where an API listens and accepts requests for a given operation or data resource.
Learn more -
A structured list of commonly asked questions and answers that address the most frequent user concerns.
Learn more -
The process of further training a pre-trained LLM model on custom data to specialize it for a specific task or style.
Learn more -
The ability of an LLM to call external functions and tools - the foundation of AI agents and chatbots that take real actions.
Learn more -
Protective mechanisms that constrain AI system behavior and prevent unwanted, harmful, or out-of-scope outputs.
Learn more -
A standard markup language that determines the structure and content of web pages.
Learn more -
A digital message that one computer sends to another to obtain or transfer data.
Learn more -
A phenomenon where an AI model generates convincingly-sounding but factually incorrect or fabricated information.
Learn more -
A combination of semantic (vector) and traditional keyword search that improves accuracy in RAG systems.
Learn more -
A universal and lightweight data format for clear storage and transfer of information between systems.
Learn more -
A type of AI trained on vast amounts of text that can understand human language and generate text.
Learn more -
An open-source framework with ready-made architecture for AI agents and integrations for any model or tool.
Learn more -
A large language model downloaded and run directly on your own computer or server without sending data to third parties.
Learn more -
Development approaches that let you build applications and automations with minimal or no coding at all.
Learn more -
An open protocol standardizing how AI models communicate with external tools, data, and services.
Learn more -
A lightweight markup language for editing plain text and converting it to formatted web content.
Learn more -
Supplementary information about data (data about data) that helps systems with organization and context.
Learn more -
An AI model that processes multiple input types at once - text, images, audio, and video.
Learn more -
The process of analyzing and converting text or data from one format into a structured form that a program can work with.
Learn more -
A text input - an instruction or question - that a user or system passes to an AI model for processing.
Learn more -
A technique where a user attempts to bypass the original instructions of an AI model using a specially crafted command.
Learn more -
A process in which AI first reads your documents and then responds to the user.
Learn more -
The maximum number of API requests that can be sent in a given time period before the server temporarily suspends access.
Learn more -
A second filter in a RAG system that refines the selection of the most relevant results from vector search.
Learn more -
A file containing a list of all important URLs on a website that helps search engines index pages more effectively.
Learn more -
Sending an LLM's answer token by token so the user sees the text appear immediately.
Learn more -
A mode where an LLM returns its answer in a predefined format (usually JSON), guaranteed to be valid and ready for downstream processing.
Learn more -
A hidden text that sets the behavior, role, and rules of an AI model before the conversation with the user begins.
Learn more -
Searching by the meaning of text rather than by exact keyword match. The foundation of modern RAG systems.
Learn more -
A parameter controlling the degree of randomness and creativity in AI model responses - from deterministic (0) to highly creative (2).
Learn more -
The basic unit of text that LLM models work with - roughly 3/4 of an English word or 2–3 characters.
Learn more -
An event or condition that automatically starts a workflow or automation.
Learn more -
A numerical representation of text meaning that allows machines to compare information by its sense.
Learn more -
A specialized database optimized for storing and semantically searching vectors - numerical representations of text.
Learn more -
An automated process of extracting data from web pages for further processing or storage.
Learn more -
An automatic notification that one system sends to another the moment a specific event occurs.
Learn more -
A sequence of steps and decisions that a system runs automatically to complete a task.
Learn more -
Prompting techniques - zero-shot gives the task without examples, few-shot includes sample inputs and outputs for better results.
Learn more -
A low-code workflow automation platform that connects hundreds of apps without complex coding.
Learn more