Let’s state this plainly before anything else: multimodal support in the WordPress AI stack is uneven right now. Image generation is real, shipped, and documented, a full builder chain exists that generates an image from a text prompt and hands you back a file you can insert straight into the media library. Vision, having a model look at an image you already have and describe or tag it, is not yet a first-class, dedicated SDK feature, there’s a workable pattern for it, but it leans on a general-purpose method rather than a purpose-built one. Audio transcription isn’t in the SDK at all today. This course teaches both halves honestly: what’s ready to build on right now, and the real workarounds for what still has gaps, without pretending either category is something it’s not.
This is Course 13 of an eighteen-course track and builds directly on two earlier
courses. Course 1: WordPress AI Foundations
introduced wp_ai_client_prompt() and the general shape of the AI Client SDK, if
you haven’t taken it, at least read its PHP AI Client SDK overview lesson first.
Course 5: LLM Integration in WordPress With the PHP AI Client SDK
goes deep on text generation, embeddings, error handling, and provider fallback, all of
which this course assumes as working knowledge rather than re-teaching. If a step here
uses generate_text() or talks about provider setup without re-explaining it, that’s
why.
By the end of this course you’ll have working plugin code that generates images straight into the media library, a documented and honest pattern for using a model to describe or tag an existing image, a real approach to audio transcription that doesn’t pretend the SDK does something it doesn’t, a bulk media analysis tool you can run against an entire library, and a clear-eyed summary of exactly where WordPress’s multimodal AI support stands today so you can build on solid ground and revisit the gaps as the SDK matures.