Auto Sourcing Relevant Pictures for Web Site Construction

[z4cth3p4c]

I’d love a way for Dyad to automatically find and insert web images that are truly relevant to each hero/section/service based on the page content. I spent several hours trying a library prompt approach (parsing section text → nouns → seeded fallbacks like LoremFlickr/Picsum). It never leaves blanks, but the relevance is hit-or-miss.

Does Dyad currently support a built-in pattern to scan the web (or a recommended hook) for content-matched images? I’m not opposed to using an optional API key if that’s the best route—I just don’t see a clear feature for it. If there’s a better way folks are using library prompts (or a lightweight integration/recipe) to get on-topic images, I’d love pointers or examples.