AI Generation
Maxi can create new content with AI — writing text and generating or editing images, on demand or in bulk. This is the capability that turns your agent from something that manages your site into something that produces for it.
Most of Maxi’s domains are about managing content you already have. AI Generation is different: it makes new material. Your agent can write copy, generate images, edit existing ones, and run any of these as large background jobs — all using the AI providers you’ve connected.
Setup first. AI Generation needs at least one provider API key configured. If you haven’t set one up, see Add and Rotate Provider API Keys — then everything below is available.
Generating text
Ask the agent to write something and it generates the text immediately. This is the engine behind tasks like drafting product descriptions, writing post intros, summarizing, or producing any copy you describe:
“Write a friendly 100-word description for this product.”
You can steer it — set the tone or role it should write in, choose which provider or model to use, and control length and creativity — but you don’t have to; sensible defaults apply. Text generation can draw on OpenAI (GPT), Anthropic (Claude), OpenRouter (a gateway to many models through one key), or a local model, depending on what you’ve configured.
Generating images
Describe an image and the agent generates it and adds it straight to your media library, ready to use:
“Generate a wide banner image of a sunlit vineyard for this page.”
You can specify size and style, and — usefully for product and design work — request a transparent background, which produces a true cut-out PNG (via OpenAI’s image model). Image generation supports OpenAI, Replicate, BFL (Black Forest Labs), and local providers.
Editing existing images
Beyond making new images, the agent can edit ones you already have — change elements, extend or alter a scene, or remove a background. For most edits you simply describe the change in words; for precise, masked inpainting (changing only a specific region) some providers take a mask. Background removal is a one-step request that returns a clean transparent PNG.
“Remove the background from this product photo.”
Doing it in bulk: batch jobs
Generating one thing at a time is fine for a one-off, but Maxi can also run batch jobs — submit many prompts at once and let them process in the background rather than waiting on each. This is how you’d handle “write descriptions for all 200 products” or “generate a thumbnail for every post in this category.” A batch returns a job you can check on:
“Generate SEO descriptions for every product missing one, as a batch.”
You can ask the agent for a job’s status (how many done, how many left, any failures) while it runs, and cancel a job that’s still in progress. Completed images from a batch are sideloaded into your media library just like single ones.
What it’s good for
- Content production at scale — descriptions, summaries, intros, variations, across many items at once.
- Visual assets — hero images, banners, product imagery, generated to your brief.
- Image cleanup — background removal and targeted edits without leaving the conversation.
- Bulk passes — applying a generation task across a whole catalog or category in one go.
In summary: AI Generation is Maxi’s content-creation capability: write text (OpenAI, Anthropic, OpenRouter, or local), generate images including transparent cut-outs (OpenAI, Replicate, BFL, or local), edit images and remove backgrounds, and run any of these as background batch jobs you can monitor and cancel. It needs a provider API key set up first; generated images land in your media library automatically.