You can attach tools to a Realtime session so the model can look up data, take actions, or call services during a live conversation. Tool configuration uses the same event surface whether your client ...

Sign in to ChatGPT with the account that owns the eligible subscription. For Pro, open Settings → Billing. For Business, open your workspace’s billing settings in ChatGPT or the Admin Console. Find ...

The conversation that this response belongs to. Items from this conversation are prepended to input_items for this response request. Input items and output items from this response are automatically ...

Configure automatic action review for an agent session to let your agents work more autonomously. Automatic review approves or denies supported approval requests as an agent works, much like the ...

Prepare a plugin's remote MCP server and optional UI for public review. Prepare a remote MCP server and its optional UI for public review as part of a plugin. This page covers remote MCP submissions.

Improve results with practical tips for fine-tuning. If you’re not getting strong results with a fine-tuned model, consider the following iterations on your process. Collect examples to target ...

Use Kubernetes as a Workload Identity Provider by exchanging a projected Kubernetes service account token for a short-lived OpenAI access token. Setting up Kubernetes This guide assumes Kubernetes ...

Creating a project gives Codex the right working environment and keeps related conversations organized. Guided practice uses prepared responses and checks, try the suggested prompts and controls. In ...

For the complete documentation index, see llms.txt. Markdown versions of documentation pages are available by appending .md to the page URL.

Azure managed identities let Azure-hosted workloads request Microsoft Entra tokens without storing long-lived secrets. In OpenAI workload identity federation, the managed identity token is the subject ...

Learn about automatic voice activity detection in the Realtime API. When VAD is enabled, the audio is chunked automatically and the Realtime API sends events to indicate when the user has started or ...

Evaluations (often called evals) test model outputs to ensure they meet your specified style and content criteria. Writing evals is an essential part of building reliable applications. Datasets, a ...