OpenAI Developers<br>Start searching API Dashboard

Search the docs<br>Search docs<br>Suggested<br>responses createreasoning_effortrealtimeprompt caching

Primary navigation

Search docs<br>Suggested<br>responses createreasoning_effortrealtimeprompt caching

Get started<br>Overview<br>Quickstart<br>Models<br>Pricing<br>SDKs and CLI OpenAI SDK<br>Agents SDK<br>OpenAI CLI

Latest: GPT-5.5<br>Prompt guidance

Core concepts<br>Text generation<br>Code generation<br>Images and vision<br>Audio and speech<br>Structured output<br>Function calling<br>Responses API<br>Using tools

Agents SDK<br>Overview<br>Quickstart<br>Agent definitions<br>Models and providers<br>Running agents<br>Sandbox agents<br>Orchestration<br>Guardrails<br>Results and state<br>Integrations and observability<br>Evaluate agent workflows<br>Voice agents<br>Agent Builder Overview<br>Node reference<br>Safety in building agents<br>ChatKit Overview<br>Customize<br>Widgets<br>Actions<br>Advanced integrations

Tools<br>Web search<br>MCP and Connectors Secure MCP Tunnel

Skills<br>Shell<br>Computer use<br>File search and retrieval File search<br>Retrieval

Tool search<br>More tools Apply Patch<br>Local shell<br>Image generation<br>Code interpreter

Run and scale<br>Conversation state<br>Background mode<br>Streaming<br>WebSocket mode<br>Webhooks<br>File inputs<br>Context management Compaction<br>Counting tokens<br>Prompt caching

Prompting Overview<br>Prompt engineering<br>Citation formatting<br>Migration guide

Reasoning Reasoning models<br>Reasoning best practices

Evaluation<br>Getting started<br>Working with evals<br>Prompt optimizer<br>External models<br>Red teaming<br>Best practices

Realtime and audio<br>Overview<br>Voice agents<br>Live translation<br>Transcription Realtime transcription<br>Speech to text

Speech generation<br>Realtime prompting guide<br>Connection methods WebRTC<br>WebSocket<br>SIP

Realtime sessions Managing conversations<br>Voice activity detection<br>Realtime with tools<br>Webhooks and server-side controls<br>Managing costs

Model optimization<br>Optimization cycle<br>Fine-tuning Supervised fine-tuning<br>Vision fine-tuning<br>Direct preference optimization<br>Reinforcement fine-tuning<br>RFT use cases<br>Best practices

Graders

Specialized models<br>Image generation<br>Video generation<br>Deep research<br>Embeddings<br>Moderation

Going live<br>Production best practices<br>Workload identity federation Overview<br>Kubernetes<br>AWS<br>Microsoft Azure<br>Google Cloud<br>GitHub Actions

Deployment checklist<br>Latency optimization Overview<br>Predicted Outputs<br>Priority processing

Cost optimization Overview<br>Batch<br>Flex processing

Accuracy optimization<br>Safety Safety best practices<br>Safety checks<br>Cybersecurity checks<br>Under 18 API Guidance

Legacy APIs<br>Assistants API Migration guide<br>Deep dive<br>Tools

Resources<br>Terms and policies<br>Changelog<br>Your data<br>Permissions<br>Rate limits<br>Admin APIs<br>Deprecations<br>MCP for deep research<br>Developer mode<br>ChatGPT Actions Introduction<br>Getting started<br>Actions library<br>Authentication<br>Production<br>Data retrieval<br>Sending files

Docs Use cases<br>Getting Started<br>Overview<br>Quickstart<br>Explore use cases<br>Migrate<br>Pricing<br>Concepts Prompting<br>Customization<br>Memories Chronicle

Sandboxing Auto-review

Subagents<br>Workflows<br>Models<br>Cyber Safety

Using Codex<br>App Overview<br>Features<br>Settings<br>Review<br>Automations<br>Worktrees<br>Local Environments<br>In-app browser<br>Chrome extension<br>Computer Use<br>Appshots<br>Commands<br>Windows<br>Troubleshooting

IDE Extension Overview<br>Features<br>Settings<br>IDE Commands<br>Slash commands

CLI Overview<br>Features<br>Command Line Options<br>Slash commands

Web Overview<br>Environments<br>Internet Access

Integrations GitHub<br>Slack<br>Linear

Codex Security Overview<br>Codex Security plugin<br>Codex Security cloud Setup<br>Improving the threat model

FAQ

Configuration<br>Config File Config Basics<br>Advanced Config<br>Config Reference<br>Environment Variables<br>Sample Config

Permissions<br>Speed<br>Rules<br>Hooks<br>AGENTS.md<br>MCP<br>Plugins Overview<br>Build plugins

Skills<br>Subagents

Administration<br>Authentication Overview<br>Access tokens

Agent approvals & security<br>Remote connections<br>Enterprise Admin Setup<br>Governance<br>Managed configuration

Windows

Automation<br>Non-interactive Mode<br>Codex SDK<br>App Server<br>MCP Server<br>GitHub Action

Learn<br>Best practices<br>Videos<br>Community<br>Blog Using skills to accelerate OSS maintenance<br>Building frontend UIs with Codex and Figma<br>View all

Cookbooks Build an Agent Improvement Loop with Traces, Evals, and Codex<br>Build iterative repair loops with Codex<br>View all

Building AI Teams

Releases<br>Changelog<br>Feature Maturity<br>Open Source

Home<br>Collections

Apps SDK Commerce Ads

Home<br>Quickstart

Core Concepts<br>MCP Apps in ChatGPT<br>MCP Server<br>UX principles<br>UI guidelines

Plan<br>Research use cases<br>Define tools<br>Design components

Build<br>Set up your server<br>Build your ChatGPT UI<br>Authenticate users<br>Manage state<br>Monetize your app<br>Examples

Deploy<br>Deploy your app<br>Connect from ChatGPT<br>Test your integration<br>Submit your app

Conversion apps<br>Restaurant reservation spec<br>Product checkout spec

Guides<br>Optimize Metadata<br>Security & Privacy<br>Troubleshooting

Resources<br>Changelog<br>App submission guidelines<br>Reference

Home

Guides<br>Get started<br>Best practices

File Upload<br>Overview<br>Products

API<br>Overview<br>Feeds<br>Products<br>Promotions

Ads Overview

Measurement<br>JavaScript Pixel<br>Conversions API<br>Supported events

Advertiser...