AI agents

An AI agent takes a goal and works through the steps for you: it plans, uses tools, edits files and reports back for your review. Compare where each agent runs, what it can access and which plans include it.

6 agentsupdated

Search AI agents by name, task or category.

Filters

Refine your search

6 agents

Newest first

ChatGPT Work

Productivity

OpenAI's agent mode in ChatGPT for turning a goal into a finished deck, report or analysis.

OpenAI Codex

Coding & development

Hand coding tasks to OpenAI's agent in the desktop app, terminal, IDE or cloud.

Claude Code

Coding & development

Delegate bug fixes, tests and migrations to a coding agent in your terminal or IDE.

How to choose an AI agent

  • Decide what it may touch

    List the files, apps and accounts the agent needs. Prefer agents that ask before changing anything important and show what they did.

  • Plan how you will review the work

    An agent can finish a long task and still get details wrong. Look for drafts, diffs or logs that let you check the result before it goes anywhere.

  • Compare usage, not only the plan price

    Agent work often draws on usage limits or credits. Check how many tasks a plan really covers and what happens when you run out.

Guides about AI agents

All guides →