Python

安装

安装支持 OpenAI Agents 的 LangSmith：

pip install "langsmith[openai-agents]"

uv add "langsmith[openai-agents]"

这将同时安装 LangSmith 库和 OpenAI Agents SDK。

环境配置

# Optional: set a project for your traces


# For LangSmith API keys linked to multiple workspaces, set the LANGSMITH_WORKSPACE_ID environment variable to specify which workspace to use.

快速入门

通过使用 OpenAIAgentsTracingProcessor class.

from agents import Agent, Runner, set_trace_processors
from langsmith.integrations.openai_agents_sdk import OpenAIAgentsTracingProcessor


async def main():
    agent = Agent(
        name="Captain Obvious",
        instructions="You are Captain Obvious, the world's most literal technical support agent.",
    )

    question = "Why is my code failing when I try to divide by zero? I keep getting this error message."
    result = await Runner.run(agent, question)
    print(result.final_output)


if __name__ == "__main__":
    set_trace_processors([OpenAIAgentsTracingProcessor()])
    asyncio.run(main())

代理的执行流程（包括跨度及其详情）会被记录到 LangSmith。

JavaScript

安装

安装 LangSmith 和 OpenAI Agents SDK：

npm install langsmith @openai/agents zod

yarn add langsmith @openai/agents zod

pnpm add langsmith @openai/agents zod

环境配置

# Optional: set a project for your traces


# For LangSmith API keys linked to multiple workspaces, set the LANGSMITH_WORKSPACE_ID environment variable to specify which workspace to use.

快速入门

在运行代理之前，向 OpenAI Agents SDK 注册 OpenAIAgentsTracingProcessor 。

setTraceProcessors([new OpenAIAgentsTracingProcessor()]);

const getWeather = tool({
  name: "get_weather",
  description: "Get the current weather for a city",
  parameters: z.object({
    city: z.string().describe("The city to get weather for"),
  }),
  execute: async ({ city }: { city: string }) => {
    return `The weather in ${city} is sunny.`;
  },
});

const agent = new Agent({
  name: "WeatherAgent",
  instructions: "You are a helpful assistant. Use the get_weather tool when asked about weather.",
  model: "gpt-5-nano",
  tools: [getWeather],
});

const result = await run(agent, "What's the weather in San Francisco?");
console.log(result.finalOutput);

生成的追踪包含根代理运行、响应跨度以及嵌套的工具调用跨度。

配置处理器

向处理器传递选项以设置 LangSmith 客户端、项目、标签、元数据或根追踪名称。

const client = new Client();
const processor = new OpenAIAgentsTracingProcessor({
  client,
  projectName: "openai-agents-demo",
  name: "Support agent workflow",
  tags: ["openai-agents"],
  metadata: {
    environment: "development",
  },
});

setTraceProcessors([processor]);

const agent = new Agent({
  name: "SupportAgent",
  instructions: "You are a concise support agent.",
  model: "gpt-5-nano",
});

const result = await run(agent, "Help me reset my password.");
console.log(result.finalOutput);

在工具中嵌套 traceable 调用

你可以在 OpenAI Agents SDK 工具处理器中使用 traceable 。LangSmith 会将这些运行嵌套在活动的工具跨度下。

setTraceProcessors([new OpenAIAgentsTracingProcessor()]);

const lookupOrder = traceable(
  async (orderId: string) => {
    return { orderId, status: "shipped" };
  },
  { name: "lookup_order" }
);

const orderStatus = tool({
  name: "order_status",
  description: "Look up the status of an order",
  parameters: z.object({
    orderId: z.string().describe("The order ID to look up"),
  }),
  execute: async ({ orderId }: { orderId: string }) => {
    return JSON.stringify(await lookupOrder(orderId));
  },
});

const agent = new Agent({
  name: "OrdersAgent",
  instructions: "Use the order_status tool to answer order questions.",
  model: "gpt-5-nano",
  tools: [orderStatus],
});

await run(agent, "Where is order 123?");

在无服务器环境中刷新追踪

在无服务器环境中追踪时，请在进程退出前刷新待处理的追踪。

const client = new Client();
const processor = new OpenAIAgentsTracingProcessor({ client });
setTraceProcessors([processor]);

try {
  const agent = new Agent({
    name: "SupportAgent",
    instructions: "You are a concise support agent.",
    model: "gpt-5-nano",
  });

  const result = await run(agent, "Help me reset my password.");
  console.log(result.finalOutput);
} finally {
  await processor.forceFlush();
}