人工智能中的代理是什么？

¥What's an agent in AI?

可以将 agent 理解为能够做出决策的 chain。当一个链按照预定的顺序调用不同的 AI 组件时，代理会使用语言模型来确定要执行的操作。

¥One way to think of an agent is as a chain that knows how to make decisions. Where a chain follows a predetermined sequence of calls to different AI components, an agent uses a language model to determine which actions to take.

代理是人工智能的一部分，可以作为决策者。它们可以与其他代理和 tools 交互。当你向客服人员发送查询时，系统会尝试选择最佳工具来解答。代理会根据你的特定查询以及配置其行为的提示进行调整。

¥Agents are the part of AI that act as decision-makers. They can interact with other agents and tools. When you send a query to an agent, it tries to choose the best tools to use to answer. Agents adapt to your specific queries, as well as the prompts that configure their behavior.

n8n 中的代理

¥Agents in n8n

n8n 提供了一个 Agent 节点，该节点可以根据你选择的设置充当不同类型的代理。有关可用代理类型的详细信息，请参阅 代理节点文档。

¥n8n provides one Agent node, which can act as different types of agent depending on the settings you choose. Refer to the Agent node documentation for details on the available agent types.

执行包含代理的工作流时，代理会运行多次。例如，它可以进行初始设置，然后运行一次调用工具，再运行一次评估工具响应并向用户反馈。

¥When you execute a workflow containing an agent, the agent runs multiple times. For example, it may do an initial setup, followed by a run to call a tool, then another run to evaluate the tool response and respond to the user.