AI startup Higgsfield went viral a few days ago thanks to its new Soul AI image generation software that lets users create stunning images with a single text prompt. That’s nothing new in the world of genAI programs, some will say. We saw ChatGPT go viral a few months ago due to OpenAI’s image generation capabilities.
Other AI chatbots can match what ChatGPT offers, with Google’s Gemini being a prime example. Google’s Veo 3 video generation software went viral after that, showing how advanced AI video-generation tools have become.
Yet Soul stood out in that increasingly crowded space thanks to a clever idea that makes Higgsfield Soul images look almost like photos taken with an iPhone. The company devised over 50 visual styles to choose from before you input your text prompt. The AI takes into account that filter-like feature along with your prompt to generate images that look like they were shot with a camera.
Any good AI image generator needs to offer users a way to edit the photos it creates. That’s the new feature from Higgsfield. Soul users can now use an Inpaint feature that lets them quickly edit specific elements of a photo.
The Higgsfield Soul Inpaint feature is now available in beta, with the company announcing it on social media.
As you can see in the clips in this post, using Inpaint is incredibly easy. It works similarly to AI video editing in other programs. You can select the items in the photos you want the AI to change, input a prompt, and let the AI take over.
Just “paint” over the items that need to be altered, and the AI will know what to replace based on your description.
Inpaint lets you upload a photo for the AI to edit. That means it might work on real photos, not just Soul creations. Higgsfield also said on X that you can generate a photo using Soul, then apply Inpaint directly by masking the object and typing the prompt.
The mini tutorial also says the feature works with Camera moves, VFX, and Avatars, which are other tools Higgsfield offers. In other words, you can create and edit AI videos with Higgsfield’s tools in addition to static images.
Again, the feature matches what many AI image generation programs already offer. We’ve had similar capabilities in Google Photos for a while. Even Apple Intelligence now lets you edit photos with AI.
But Soul might take it to another level. Higgsfield pitches its AI tools as programs creators can use to quickly make content for social media. You can change outfits, hairstyles, and accessories “in seconds” with Inpaint, instead of conducting lengthy photo shoots where your subject has to change outfits multiple times.
The tools are also great if you want to “generate memes, stories, or entire faceless channels with precise edits.”
I’ve said before that using tools like Google Photos to create fake memories makes me uneasy. I’m not a big fan of how easily regular users can “Photoshop” images with AI, especially since most AI tools don’t include visible watermarks that show an image was generated or edited with AI.
Higgsfield Soul and Inpaint make it incredibly easy to create and edit content. It might be even easier than using ChatGPT, Google Photos, Adobe Firefly, or the many other genAI tools out there. I just wish these images always included a visible watermark so people could immediately tell they’re looking at AI-generated content.
Some of the samples Higgsfield shared on X include a “Made in Higgsfield” watermark, but it’s unclear whether that’s a permanent feature or just something added for those specific clips.