Unified TypeScript library for streaming, structured outputs, and agent loops across AI providers
Full evaluationGlossary · Updated Sep 2, 2026
AI agent
A system that uses a model to plan and take actions with tools over multiple steps toward a goal, not just answer once.
Definition
An AI agent loops: it reads a goal, decides on an action such as searching, calling an API, editing a file, or asking a question, observes the result, and continues until done. Agents range from support bots that look up orders to coding agents that implement features. Reliability depends on tool design, guardrails, and evaluation.
Why it matters when choosing a tool
Agents can save real work but also act wrongly at scale, so tools marketed as agents should show what actions they can take, how they escalate, and how you audit them.
Where you will meet it
AI Agent & Chatbot Builders, AI Automation & Workflows, AI Coding & Development
Related terms
Agentic workflow · Tool use (function calling) · Guardrails · Model Context Protocol (MCP)
Tools where this matters
Reviewed products in the related categories
Prototyping environment and developer API platform for Google's Gemini models
Full evaluation
A source-available workflow automation platform with native AI agent nodes, self-hostable or cloud.
Full evaluation
AI writing, search, and task assistance inside a connected workspace and knowledge base.
Full evaluation
AI-assisted automation across a large app ecosystem, tables, agents, and business workflows.
Full evaluation
Anthropic's agentic coding tool that works in the terminal, IDE, desktop app, and browser to plan and execute multi-step changes.
Full evaluation