AI Social Image
What stalls a post is rarely taste — it's opening another tool and working out the size again
You've written the post; the last step is the picture — and that step often takes longer than the writing. Stock images mean checking licences, templates mean fighting the text, drawing it yourself means time you don't have. Worse, every platform wants a different shape: 3:4 for a vertical note, a wide banner for a newsletter header, square tiles for a grid. This page folds all three into one: describe the post or the topic, and it lays out an image at the size you need, with text and background on their own layers.
The text stays editable afterwards — the most important thing for a post image, because copy often gets polished again after the image is done. Changing text, colour or position only touches that layer. The AI creation step runs in the cloud (prompt only), with no sign-in first; the hands-on parts — layout, adding text, export — all run on your device and work offline.
Key features
Laid out at the platform's size
Vertical note, newsletter header, square post, grid tile — name the platform. The size is fixed from the start, so nothing gets cropped off at the end.
Text and background are separate layers
Polish the copy and you change only the text layer; the background stays. Or the other way round: keep the words and try a different background underneath.
A consistent set, not one-offs
A single post often needs several images. Ask for the whole set at once and it keeps one palette and one type treatment throughout, zipped at the end.
Your own photo works too
Drag a photo in (it catches the drop anywhere on the page) and have it lay out text and decoration on top — at that point the assistant takes a thumbnail screenshot and sends it to the cloud. A photo you took yourself often fits the content better than a purely generated image.
How to use
- 1 Open the editorClick the button below to create a canvas and open the AI assistant chat — no material to prepare first.
- 2 Say the content and the platform"An image for this note: 'Spent the weekend up in the hills', quiet tone, portrait" — the content sets the words, the platform sets the size.
- 3 Adjust until it fitsPolished the copy? Say "use this line instead". Tone off? Say "lighter background". Or adjust it yourself.
- 4 ExportPNG / JPG to publish now, PSD to change it next time. No watermark, no export limit.
FAQ
- Is there any copyright risk in the generated image?
- Exports carry no watermark and we claim no rights over them. One thing to watch: if your prompt asks it to imitate a specific brand, celebrity or copyrighted work, the output may touch someone else's rights — that judgement sits with you. Using your own words and your own photo as the base is the safest route.
- Do I need an account for social images? Does my image get sent out?
- No account. The AI generation step runs in the cloud and sends only the prompt — no sign-in, free credits every day; when the canvas already has content, the assistant takes a thumbnail screenshot to look at it. Layout, on the other hand, is entirely on your device: dragging in your own photo, adding text, adjusting colour, exporting — that whole path is unlimited and works offline.
- Can a set of images share one style?
- Yes. Say up front how many and on what theme, and it works through them with one palette and one type treatment. You can also make the first one, then say "three more like this, with the text changed to…".
- How do I make grid tiles?
- Make one large square image first, then use the grid-split tool to cut it into 9 — that step runs entirely on your device, with no upload and no sign-in. The numbering matches the order you'll post them in.
Try AI Social Image now
Creates a canvas and opens the AI assistant chat · AI generation runs in the cloud (prompt only) · your own layout and export all run on your device