What you can make
Images
- Cover photos and feature openers in portrait or landscape shapes, with room left for a headline.
- Spot illustrations and product shots. OpenAI models can return a transparent background, so the picture sits on your page color.
- Background textures quiet enough to set text on.
- Variations of a picture you already have. Add it as a reference image and describe what should change.
Copy
- Headlines, decks, pull quotes and captions: several options at once, each with its own Place button.
- Body text at the length you ask for, and flyer or poster text with headline, details and a call to action on separate lines.
- Rewrite your own text: shorten it, fit it to a word count so it stops overflowing its frame, expand it, change the tone, proofread it or translate it.
How results get into DesignCraft
The browser build of DesignCraft has no scripting interface, but it does accept files dropped on its canvas. Place in DesignCraft hands the image or text to the editor in exactly that way, so it lands where a dragged file would:
| You have… | An image goes | Text goes |
|---|---|---|
| The Type tool’s cursor in a text frame | a new frame on the first spread | in at the cursor, replacing selected text |
| An empty frame selected | into that frame, filling it in proportion | into that frame, which becomes a text frame |
| A picture frame selected | into that frame, replacing its picture | a new text frame inside page 1’s margins |
| A text frame selected | a new frame on the first spread | into that frame, replacing its text |
| Nothing selected | a new frame at the first spread’s top-left margin | a new text frame inside page 1’s margins |
So the quickest way to fill a picture box is to draw it with the Rectangle Frame tool (F), leave it selected and press Place. Text arrives as plain text with one paragraph per line; apply your paragraph styles to it afterwards. You can also Save any result and place it later with File › Place….
Step by step
- In the editor, open a document (or create one with New file), then click AI tools in the top bar. The panel opens on the right and the canvas makes room.
- Open the settings (gear icon), choose a provider and paste your API key. Each provider has an image model and a text model; the defaults work, and Load models shows what your key can use.
- Images: pick an idea such as Cover photo or Spot illustration, fill in the subject, choose a shape and press Generate image (or ⌘/Ctrl + Enter).
- Copy: choose Headlines, Body text, Flyer or another kind, describe what the page is about and press Write. For a rewrite, paste the text from your frame.
- Select the frame you want to fill in DesignCraft, then press Place in DesignCraft on the result.
Supported providers
| Provider | Default image model | Default text model | Get a key |
|---|---|---|---|
| OpenAI | gpt-image-2.5-flare | gpt-6.1-sol | platform.openai.com |
| Google Gemini | gemini-3.1-flash-image | gemini-3.8-flash | aistudio.google.com |
| OpenRouter | google/gemini-3.1-flash-image | google/gemini-3.8-flash | openrouter.ai |
| Custom (OpenAI-compatible) | any model your API serves | any model your API serves | your own endpoint |
OpenRouter gives one key access to many image models, including Gemini, GPT Image, FLUX, Seedream, Qwen Image and Recraft, and to text models from most labs. The custom option works with any service that implements OpenAI’s /images/generations and /chat/completions endpoints and accepts browser requests.
Privacy: what goes where
| Data | Where it goes |
|---|---|
| Your API key | Stays in your browser. Sent only to the provider, in the request header. |
| Your prompt, brief, pasted text and any reference image | Sent from your browser straight to the provider you picked. |
| Your DesignCraft document | Never leaves your device. The panel can’t read it. |
| designcraft.run servers | Serve the static editor files. They receive none of the above. |
The provider’s own terms and data policy apply to what you send them. Reference images are scaled to 2048 pixels on the long side before upload to keep requests fast and within provider limits.
Image size and print
OpenAI image models return 1024 × 1024, 1536 × 1024 or 1024 × 1536 pixels, depending on the shape you pick, and most other models are in the same range. That is plenty for screens, posts and small printed pictures, but a full-page print photo usually wants more pixels, so check the result at its printed size before you send a PDF to the printer.
FAQ
Do I have to pay to use the AI tools?
You pay your AI provider for what you use, at their normal API prices. designcraft.run charges nothing and adds no markup, because the request never passes through us.
Where is my API key stored?
In your browser only. By default it lasts for the current tab (sessionStorage). Tick “Remember my key on this device” to keep it in localStorage instead, and use “Forget all keys” in the panel settings to delete it.
Where does a placed image go?
Into the empty or picture frame you have selected, scaled to fill it in proportion. Otherwise DesignCraft makes a new frame at the top-left margin of the first spread. A document has to be open, or DesignCraft cancels the import and says so in its status bar.
Where does placed text go?
If you are typing in a text frame with the Type tool, it goes in at the cursor and replaces any selected text. If a frame is selected, it replaces that frame’s text. Otherwise it becomes a new text frame inside the margins of page 1.
Can the AI read or rearrange my layout?
No. The browser version of DesignCraft has no way for another program to read the document, so the panel only adds things to it. To rewrite text that is already on the page, copy it from DesignCraft and paste it into the panel. To have an agent work on the document itself, see the MCP server for the desktop app.
Which model should I pick?
For images, gpt-image-2.5-flare (OpenAI) and gemini-3.1-flash-image (Google) are quick everyday choices. For copy, any current chat model writes headlines and captions well; gemini-3.8-flash and gpt-6.1-sol are the defaults. “Load models” in the panel lists everything your key can use.
Why does a request fail with a CORS or network error?
The custom provider option needs an API that accepts requests from browsers. OpenAI, Google Gemini and OpenRouter all do. Some self-hosted or regional gateways don’t; use one of the built-in providers instead.