Images and the Media Library

Every image in a project — generated or uploaded — lives in the Media Library, and the same generator is available wherever you need an image: the Media Library page, the Canvas Image tool and each campaign or calendar post. Images are stored durably in your project, so anything you generate can be browsed, reused and deleted later.

Where you work with images

• The Media Library shows every image for the active project, newest first, with the generator on top. You can paste content to match, generate from a prompt, or upload your own images.
• The Canvas Image tool generates an image that matches the document you are writing, pre-fed with its title and body. Inserting adds the image to the end of the document as a review-only edit, like any other suggestion.
• Campaign and calendar posts have a per-post Generate from post button that derives an image from that post’s own copy, plus an Upload option. An image is optional on most platforms and required for Instagram.

Matching your content

The generator has two modes: Match this content and Describe it myself. In match mode, WordPresto reads your real copy and writes one image prompt that represents that specific piece — its subject, mood and setting — and never invents brand names, logos, statistics, in-image text or claims that are not in your copy. Writing that prompt is a small metered step, so it uses credits; describing the image yourself does not. Generation only ever runs when you press Generate.

Models and options

You can choose the model that generates the image. The default is tuned for text-heavy editorial layouts and infographics, with alternatives for fast low-cost generation, high-realism hero art, top prompt adherence, and a distinct stylistic look. Each model shows only the controls it really supports — shape and size, number of images, format and compression, render quality, safety-filter strictness, steps and guidance, seed, or a negative prompt — so you are never offered an option the model would ignore.

All models accept optional reference images: attach up to sixteen from an upload or from the project’s library, and the model edits or takes guidance from them. That is the path for editing an existing image or keeping a product or character consistent across a series.

Heavier models run in the background

Most models return their image straight away. A heavy model keeps generating in the background: the request is accepted immediately and the finished image drops into your library automatically, even if you navigate away. If it is still working after a few minutes you will see “Still generating — it will appear in your library shortly”, which means exactly that.

Honest failures

If something goes wrong you see the real reason, not a spinner: a safety filter blocked the prompt (often a false positive — rephrase, relax the model’s safety option or switch model), the model timed out (try a smaller size, fewer images or a faster model), or there were not enough credits to write the content-matched prompt. Nothing is charged for work that did not complete and store.