Skip to content

Try out the Meta image maker – Imagine

Meta has joined the ‘text to image’ AI race with Imagine. Here’s how to use it to make interesting pictures quickly and bypass its current limits.

There are so many AI tools out there, in particular ‘text to image’ makers, Microsoft has ‘Generate Image’ in Designer and a separate Image Creator as well as the same tech in CoPilot. All those are based on Dall-E, part of ChatGPT.

Meta (aka Facebook) has released their own tool ‘Imagine’ which is interesting for two reasons.

  • It uses quite different AI technology (codenamed ‘Emu’) than the Microsoft/ChatGPT system. That means you can expect noticeably different images from Imagine.
  • Imagine is FAST.  Four images appear on the screen a lot quicker than other systems.
Our standard AI test prompt – “Koala driving a red London bus”

Starting Imagine

Go to https://imagine.meta.com/ and login, most likely with a Facebook account.

At present, Imagine is limited to users in the USA but we used a VPN connection via a US server to bypass that.

Type in a description plus, optionally, a style to get some images.

Click on each thumbnail to see the full-size image.

Like other AI systems, you can add an artistic medium, style or artist. Maybe a watercolor …

Or in the style of Salvador Dali (with a lobster in the image, we had to try Dali) …

Save what you like, right away

Unlike the Microsoft and OpenAI tools, Imagine does NOT save all your generated images in a library.  Perhaps that’s a future extra?

For the moment, save any image you like because there’s no way to recall past images.

Download

Click on the menu button top right of the image and choose Download.

That copies a JPG/JPEG file to your computer, 1280 x 1280 at 96dpi, 24 bit color. The image name is the text prompt.

Any Imagine JPG can be inserted into a Word doc, PowerPoint slide, Outlook email or in Microsoft Designer.

Watermark

Bottom left of each image is a visible watermark.

Remove that by cropping the image (if possible) or a smart erase / spot healing tool.

Meta says that an ‘invisible’ watermark will be added in the future.

We’d like all the AI image makers to add some basic details to the downloaded image properties. AI generated images should have the maker, text description and other details in the file properties. These are simple extras that would benefit users.

Why a mouse riding a lobster?

Why did we choose ‘mouse riding a red lobster’ for the tests of Imagine?

It was from a ‘Lobster riding mouse’ Christmas Card circa 1880 part of a collection shown in Shock of the old: 11 murderous and macabre Victorian Christmas cards

Designer home page changes are a welcome improvement
Designer gets a big AI image boost
What is Dall-E and what it’ll do for Microsoft 365 customers

About this author

Office-Watch.com

Office Watch is the independent source of Microsoft Office news, tips and help since 1996. Don't miss our famous free newsletter.