AI API

The AI API is an OpenAI-compatible HTTP service for chat completions, image generation, text-to-speech, transcription, and content tooling. One API key reaches models from Google, Anthropic, OpenAI, Qwen, and others. You pick a model by its ID, such as google/gemini-3.8-flash.

To start, create an API key in the WP Engine User Portal. Then point a client that speaks the OpenAI Chat Completions API at https://api.ai.wpengine.com/v1 and make your first call. The Responses, Embeddings, and legacy Completions APIs aren’t available.

Beta

This documentation covers the AI beta. Endpoints, model availability, and rate limits may change before general availability.

Call the API directlyBrowse the model catalog, create a key, and make your first chat completion.

Use AI inside WordPressFor plugins and themes, the AI Client bundled with WordPress handles auth and provider routing for you.

Connect an AI agentGive Claude Code, Claude Desktop, or any Model Context Protocol (MCP) client access to your WordPress sites.

---

Before your first call

• Authentication — create, scope, and rotate a key
• Site and WordPress account connections — why you still can't write to a registered site

After your first call

• Stream a chat completion — parse Server-Sent Events in JavaScript and PHP
• Errors — response shape and common status codes
• Rate limits — quotas and 429 handling