Adobe is introducing a new wave of AI-driven capabilities to its experimental iOS camera app, Project Indigo, which first launched last year. The app already offered professional controls, multi-frame super-resolution, and various capture modes. Now, it’s adding features that leverage large language models (LLMs) to analyze photos and offer editing recommendations. Additional AI tools include advanced object removal, depth-of-field generation, and the ability to apply different artistic styles to images.

Marc Levoy, who leads Adobe’s project and previously developed the camera technology for Google Pixel phones, noted that most generative AI tools rely on prompt-based editing. Finding the perfect prompt to tweak a photo or achieve the desired result can often be tricky. To address this, most of the new experimental AI features are designed as buttons that produce more predictable outcomes. For example, Google’s camera coach feature for Pixel phones, released last year, provided only basic framing suggestions. In contrast, Project Indigo’s features offer detailed, descriptive guidance that can teach users photography techniques, even if they disagree with the AI’s assessment. Two key features fall into this category. First, the photo critique delivers a “professional” evaluation of framing, lighting, colors, and emotional impact.

The second feature provides capture and editing suggestions, offering tips on how to reshoot a photo by adjusting framing, exposure, and objects in the viewfinder. For instance, the app once advised me to remove a hexagonal white object from the frame of a photo I took. A separate section then explains how to improve an existing photo using Adobe Lightroom controls.

Object removal has been available in photography apps like Apple Photos, Google Photos, and Adobe Photoshop for years. However, these features typically require users to manually circle or select an object by drawing on the screen, which isn’t always precise. Project Indigo’s new feature introduces toggles for removing specific elements from an image, such as people in the background, trash and trash cans, wires and poles, fences, vehicles, and other clutter. Users can also describe a custom object for removal. The results are impressive—the app successfully removed a friend and the object he was holding from the background without leaving any strange artifacts.

The app also uses AI to generate depth of field, simulating a blurred background in photos. With this update, Adobe is experimenting with a style transfer feature that transforms images into effects like watercolor, pen and ink, ink line with color wash, monochromatic, and backlit subject. Some of these styles evoke the early days of the Prisma app. While the outputs look refined thanks to advanced models, style transfer isn’t a novel or particularly useful feature. Despite Levoy’s critique of prompt-based implementations, Project Indigo includes its own feature that allows users to describe their desired edit. This can be used to perform editing workflows not available in preset tools, though it also risks leading to sloppy results. The company is currently using Google’s Gemini-based Nano Banana for these features, but remains open to swapping in other models, including its own Adobe Firefly model. These capabilities, grouped under the AI playground tab, are still in testing and only available to select users. They may never reach a broader audience, but it’s encouraging to see features that help people understand photography nuances, rather than simply generating more slop.