Skip to content
GlossaryConceptUpdated May 2026

Agent

noun · also: llm, tool-use, rag

What is agent?

An AI agent is an LLM that can take actions in the world — not just generate text, but call tools, browse the web, and chain steps to complete tasks.

Definition

Full definition of agent

Agents work by giving the LLM a set of tools (functions it can call), letting it decide which to use, executing the call, feeding results back, and looping until done. Useful for tasks that require multiple steps or external data lookup. Tiny Command's TinyAgents builds on this pattern with a visual builder. Examples: research bot that searches web + summarizes, support bot that looks up CRM + drafts reply.

In practice

Agent examples

Agent flow
User: 'Refund the last charge for Acme.' → Agent calls Stripe Search → Calls Stripe Refund → Reports result.
Used by

Apps that exemplify agent

See agent in action across real integrations.

FAQ

Common questions about agent

Are agents safe to run autonomously?
Not for consequential actions. Add human-approval gates for irreversible operations (refunds, deletes, sends). Use Tiny Command's approval-message pattern.
What's the difference between an agent and a workflow?
Workflows execute fixed steps. Agents pick which steps to run based on context. Agents are more flexible but harder to predict and debug.