What is an AI agent?
An AI agent is a system that uses a language model to plan and take actions toward a goal, calling tools, reading results and deciding the next step in a loop, rather than just answering a single question.
The difference between a chatbot and an agent is autonomy over multiple steps. Given a goal like "find the bug and fix it", an agent breaks it into steps, uses tools (a shell, a browser, an MCP server) to gather information and make changes, checks the outcome, and repeats until the goal is met or it gets stuck. The model is the reasoning engine; the agent is the loop around it.
Agents range from focused coding assistants that edit your repo and open pull requests, to research agents that crawl dozens of sources and write cited reports, to general workspaces that can run shell commands and browse the web unattended.
The strongest agents pair a capable model with good tools and tight feedback: small, reviewable steps beat one giant leap. Many of the agents here are open source and install in a couple of commands.
Popular agents
All Agents →Type
- pydantic-aiAI Agent Framework, the Pydantic wayAgentsJul 2026
- adk-pythonAn open-source, code-first Python toolkit for building, evaluating, and deploying sophisticated AI agents with flexibility and control.AgentsJul 2026
- mini-swe-agentThe 100 line AI agent that solves GitHub issues or helps you in your command line. Radically simple, no huge configs, no giant monorepo—but scores >74% on SWE-bench verified!AgentsJul 2026
- ScienceClaw🔬🦞 A self-evolving AI research colleague for scientists. 285 skills, zero hallucination, persistent memory.AgentsJun 2026
- dashA self-learning data agent built with systems engineering principles. It grounds answers in 6 layers of context and improves with every query.AgentsJul 2026
- bedrock-agentcore-sdk-pythonPython SDK for transforming any AI agent into a production-ready application. Framework-agnostic primitives for runtime, memory, authentication, and tools with AWS-managed infrastructure.AgentsJul 2026
FAQ
What's the difference between an agent and an MCP server?
The agent is the decision-maker running the loop; MCP servers are the tools it calls. An agent becomes far more capable when you connect the right servers.
Are AI agents safe to run unattended?
Give them scoped permissions and review their changes. Coding agents that work in git let you inspect every diff before it lands.