Agents
The word everyone uses and almost nobody defines precisely
What are agents
Think of the difference between a calculator and a spreadsheet. A calculator does what you ask, once. A spreadsheet watches for changes and recalculates automatically. An AI agent is closer to the spreadsheet — it takes a goal, decides what steps to take, executes them, observes the results, and repeats until the goal is reached.
An agent uses a language model to decide which action to take next. The actions can include calling APIs, querying databases, running code, or browsing the web. The model is the brain. The tools are the hands. The agent loop is the decision cycle that connects them.
Why it matters
Agents are the architecture behind Cursor, Devin, Claude Code, and every AI coding assistant that can edit files, run tests, and iterate. The distinction matters because agents consume far more tokens than single-call chatbots — multiple model calls per task, each with tool-use overhead. Understanding agent architecture is understanding your cost structure. sourc.dev tracks agent tools at [/tools](/tools).
Verified March 2026 · Source: sourc.dev entity data