# Glossary #### **AI Agent** An AI-powered system that can interpret requests, make decisions, and perform actions using large language models (LLMs). AI agents can hold context, respond intelligently, and interact with tools and data sources. *** #### **AI Chain** A structured sequence of calls to LLMs or external resources. AI chains execute logic step-by-step but **do not retain memory** across executions. Use AI agents for persistent conversations. *** #### **AI Embedding** A numerical vector representation of data that captures its meaning and relationships. Used to power search, recommendation, and similarity logic in AI. Stored and queried using **vector databases**. *** #### **AI Memory** Enables stateful conversations with AI by remembering past interactions. Essential for multi-step or context-aware tasks. Available in AI agents, not AI chains. *** #### **AI Tool** A functional resource the AI can call during execution. Tools can include APIs, databases, or logic modules—used for extending an agent’s capabilities. *** #### **AI Vector Store** A database designed to store and retrieve **embeddings**—useful for tasks like semantic search or similarity comparison. *** #### **API (Application Programming Interface)** A programmatic interface to a service or application. APIs allow workflows to send and receive data without human interaction. *** #### **Canvas** The visual interface where you build workflows. Nodes are added, connected, and configured directly on the canvas. *** #### **Cluster Node** A modular component composed of a **root node** and one or more **sub nodes**. Used for grouped or extended logic blocks. *** #### **Credential** Secure authentication information (like API keys or tokens) that allows workflows to connect with external apps or services. *** #### **Data Pinning** Freezes the output of a node for testing. Prevents repeated external calls and lets you debug with consistent, mockable data. Ignored during live (production) runs. *** #### **Editor** The full visual interface for creating workflows. Includes the canvas, node settings, execution logs, and controls for variables, credentials, templates, and more. *** #### **Entitlement** Access rights tied to a subscription or plan. May be used to unlock premium or restricted features within Autom8 AI. *** #### **Evaluation** The process of comparing execution outcomes over time. Used to measure workflow performance, especially for AI-driven logic. *** #### **Expression** A dynamic value written in JavaScript-like syntax to pull in data from other nodes, environment variables, or runtime context. *** #### **LangChain** A popular open-source framework for building AI apps using LLMs. It connects prompts, models, memory, and tools to build complex AI workflows. *** #### **LLM (Large Language Model)** An advanced AI model trained on large datasets to understand and generate human-like text. Powers most AI agents and chains. *** #### **Node** A functional unit in a workflow that performs an action, such as fetching data, sending an email, or transforming input. Nodes can act as triggers, processors, or integrations. *** #### **Project** A way to group workflows, credentials, and resources for better organization and collaboration across teams. *** #### **Root Node** The primary logic unit in a **cluster node**, extended by one or more sub nodes. *** #### **Sub Node** A secondary component that extends the behavior of a root node, typically used in grouped logic structures. *** #### **Template** A ready-to-use workflow created to solve common use cases. Templates can be customized with your own data, services, and logic. *** #### **Trigger Node** A node that starts a workflow based on specific events or schedules. Required for all production workflows. *** #### **Workflow** A visual automation composed of connected nodes. Begins with a trigger and flows step-by-step to complete a task or process.