- 入门指南
- Studio Web 中的 UiPath 智能体
- UiPath 编码智能体
- 使用编码智能体构建
Agent execution in Agents: a single execution model that runs low-code and coded agents consistently, whether triggered from Debug or Orchestrator.
All agents run through the same execution engine, regardless of how they are built. This provides a single runtime model that ensures agents behave consistently whether they are created using low-code tools or written directly in code.
At a high level, this shared execution model provides:
- 一个适用于所有智能体的执行引擎
- 一个运行时合同,用于定义智能体的运行方式
- 一个运营模型,用于调试、监控和恢复执行
Agents are represented using a standardized definition that includes prompts, tools, inputs, outputs, and settings. At runtime, this definition is interpreted and executed by the same underlying engine.
这促成了一种“两全其美”的方法:
- 低代码用户可以使用可视化工具快速构建和迭代
- 专业开发人员可以在需要时扩展智能体或将智能体转换为代码
- 企业受益于一致的监管、安全性和可观察性
这消除了低代码智能体和编码智能体之间的碎片化,避免了在不同开发风格之间切换时需要重新构建智能体。
Agent execution model
当智能体运行时:
- 智能体定义被打包为一个自包含的执行合同。
- 运行时会在一个安全的无服务器环境中执行智能体。
- 长时间运行的步骤(例如人机回圈操作或外部流程)可能会暂停执行。
- 在所需操作完成后,智能体会自动恢复。
- 通过追踪和日志捕获执行详细信息。
All of this happens using the same engine, regardless of whether the run is triggered from Debug or from Orchestrator. From a user interface perspective, this execution model is largely invisible.
从 Agent Builder 和 Orchestrator 中:
- 选择调试以测试智能体。
- 您将智能体作为解决方案发布到 UiPath 租户,并作为 Orchestrator 作业运行 — 通过 API、Maestro 节点或流程活动触发。
- 您将查看输入、输出、工具调用和追踪信息。
- 您将看到智能体运行会经历熟悉的状态:运行中、待处理、已暂停、已完成或失败。
追踪
Because all agents run on the same engine, traces have a consistent structure regardless of whether the agent was built with low-code tools or written in code. The traces UI surfaces:
- LLM 调用和响应
- 工具调用和结果
- 错误和防护栏操作