A practical guide to building agents

# A practical guide to building agents | OpenAI. Log inTry ChatGPT(opens in a new window). Try ChatGPT(opens in a new window)Login. # A practical guide to building agents. Try ChatGPT(opens in a new window)Contact sales. Advances in reasoning, multimodality, and tool use have unlocked a new category of LLM-powered systems known as **agents**. This guide is designed for product and engineering teams exploring how to build their first agents, distilling insights from numerous customer deployments into practical and actionable best practices. It includes frameworks for identifying promising use cases, clear patterns for designing agent logic and orchestration, and best practices to ensure your agents run safely, predictably, and effectively. After reading this guide, you’ll have the foundational knowledge you need to confidently start building your first agent. 2. It has access to various tools to interact with external systems—both to gather context and to take actions—and dynamically selects the appropriate tools depending on the workflow’s current state, always operating within clearly defined guardrails.

AI Agent行业深度分析

1. Agents SDK | OpenAI API

Build agents in code with the OpenAI Agents SDK and grow into more advanced runtime patterns as needed. * Use the **Agents SDK** pages when your application owns orchestration, tool execution, approvals, and state. * Use **Agent Builder** only when you specifically want the hosted workflow editor and ChatKit path. | Build a code-first agent app | Quickstart | This is the shortest path to a working SDK integration. | Inspect and improve runs | Integrations and observability and evaluate agent workflows | Use traces for debugging first, then move into evaluation loops. | Build a voice-first workflow | Voice agents | Voice is still an SDK-first path because Agent Builder doesn’t support it. Use the SDK track when your server owns orchestration, tool execution, state, and approvals. ## Use Agent Builder for the hosted workflow path. Use Agent Builder when you want OpenAI-hosted workflow creation, publishing, and ChatKit deployment. Voice agents are an exception: they live in the SDK track because Agent Builder doesn’t currently support voice workflows.

查看完整报道 →

2. OpenAI Agents SDK – GitHub Pages

from agents import Agent, Runner, function_tool, ModelSettings from agents import Agent, Runner, function_tool, ModelSettings @function_tool @function_tooldef get_weather(city: str) -> str: def get_weather(city: str) -> str: “””Returns weather info for the specified city.””” “””Returns weather info for the specified city.””” return f”The weather in {city} is sunny” return f “The weather in {city} is sunny” agent = Agent( agent = Agent( name=”Weather Agent”, name = “Weather Agent”, instructions=”Retrieve weather details.”, instructions =”Retrieve weather details.”, tools=[get_weather], tools =[get_weather], tool_use_behavior=”stop_on_first_tool” tool_use_behavior = “stop_on_first_tool”) ). from agents import Agent, Runner, function_tool, FunctionToolResult, RunContextWrapper from agents import Agent, Runner, function_tool, FunctionToolResult, RunContextWrapperfrom agents.agent import ToolsToFinalOutputResult from agents.agent import ToolsToFinalOutputResultfrom typing import List, Any from typing import List, Any @function_tool @function_tooldef get_weather(city: str) -> str: def get_weather(city: str) -> str: “””Returns weather info for the specified city.””” “””Returns weather info for the specified city.””” return f”The weather in {city} is sunny” return f “The weather in {city} is sunny” def custom_tool_handler( def custom_tool_handler( context: RunContextWrapper[Any], context: RunContextWrapper[Any], tool_results: List[FunctionToolResult] tool_results: List[FunctionToolResult]) -> ToolsToFinalOutputResult: ) -> ToolsToFinalOutputResult: “””Processes tool results to decide final output.””” “””Processes tool results to decide final output.””” for result in tool_results: for result in tool_results: if result.output and “sunny” in result.output: if r

查看完整报道 →

3. Solutions for agentic workflows – OpenAI

Image 1: Workflow interface showing a connected sequence starting with a green “Start” button leading to a blue “Categorize Agent” step on a soft blue and orange gradient background. Explore OpenAI products that help you use, build, and scale agentic AI. Image 2: ChatGPT interface with browser tabs showing Aerion shoe pages reopened after a user request, with a note confirming four tabs opened and an option to pull more from the last month. * Create customer-facing agentic experiences in your products or specialized internal agents for your team. See how sales, marketing, IT, and product teams use workspace agents to move faster and automate recurring tasks. Image 6: Interface labeled “Evaluate traces” displaying an API log with customer complaint text about a canceled flight and the AI-generated response apologizing for the issue. Connect ChatGPT and agents to your organization’s tools using apps and MCP, or develop custom integrations for your systems. * Explore ChatGPT(opens in a new window). * API log in(opens in a new window).

查看完整报道 →