Turn your doodles into “photos” with StabilityAI’s latest AI tool

Tips & Techniques

Stability AI has launched Stable Doodle, a sketch-to-image tool that lets you convert your drawings, even very basic ones, into a photorealistic image, artwork, and more. I gave it a quick spin, so let’s dive in and see how it works.

What’s Stable Doodle and how to use it?

This new tool from Clipdrop by Stability AI is great for us who suck at drawing, I guess. StablilityAI writes that it has “the ability to dramatically enhance a range of industries, from education to creative design, to fashion and the arts.” Or shall I say, replace people who have devoted years of education and practice to become designers and artists? But I won’t get so realistic dark in this article. Moving on!

Just like Adobe Firefly, which I tested just yesterday, Stable Doodle offers a user-friendly, intuitive interface. You can create a basic sketch with a mouse, choose an art style, write a textual prompt, and click “generate.” You’ll get four results, and from there on, you can refine them and play with it, apply different styles, change your prompts, etc.

Examples

Here are a few examples I made, real quick. First, I made a “castle with two towers and a wall with a river surrounding it and a forest behind it.” I chose “photographic” stale, can’t say they nailed it. But then again, I don’t know how to draw. 🙂

Next I tried “a flock of birds flying above the city towards sunset” in “photographic” and “cinematic” style. Very Hitchcockian results, I like them.

How it’s made

Stable Doodle combines the image-generating technology of Stability AI’s Stable Diffusion XL with the powerful T2I-Adapter. “T2I-Adapter is a condition control solution developed by Tencent ARC (license), StabilityAI explains, and “it allows for precise control over AI image generation.”

Image credits: StabilityAI

“By adding trainable parameters to existing large diffusion models, the T2I-Adapter enables inclusion of additional input conditions, such as sketches, segmentation maps, or key poses. This framework supports multiple models for input guidance simultaneously, offering enhanced control over the generation process. For the Stable Doodle use case, the T2I-Adapter provides supplementary guidance to the pre-trained text-to-image model (SDXL) to enable it to understand outlines of sketches and generate images based on prompts combined with the outlines defined by the model.”

Limitations disclaimer

“Although Stable Doodle demonstrates impressive capabilities, it is important to recognize its inherent limitations,” StabilityAI warns.

“Stable Doodle uses algorithms to analyze the outline of an image to generate a visually pleasing and coherent result. The final output is dependent on the initial drawing and description provided by the user and the accuracy of the tool may vary depending on the complexity of the scene.

Users should exercise caution when relying solely on Stable Doodle for critical applications. As with all Clipdrop tools, users of Stable Doodle must comply with the Clipdrop General Terms and Conditions.”

Availability and terms

You can try Stable Doodle on the Clipdrop website and as an app (iOS and Google Play). You can start experimenting for free without a login, but there’s a daily limit of only one attempt unless you sign up, or wait one hour.

When you sign up with your email, go ahead and have fun. A free account has a limit on the number of creations, but I haven’t found the exact number (I’ll update if I find it). But if you want to go limitless, it will cost you $7 a month and cover all StabilityAI’s tools.

Leave a Reply

Your email address will not be published. Required fields are marked *