Adobe's Project Indigo adds AI photo critique and object removal

Adobe is expanding Project Indigo, an experimental iOS camera app launched last year, with several new AI-powered capabilities designed to help users critique and improve their photographs. The app previously offered pro controls, multi-frame super-resolution, and different capture modes. The newest additions include LLM-powered photo critique, advanced object removal, depth of field generation, and style transfer.

Marc Levoy, who heads the project and previously led camera development for Google's Pixel phones, explained the design rationale behind these features. He noted that most generative AI tools rely on open-ended prompts, which can be unpredictable—finding the exact phrasing to achieve the desired result often proves difficult. To address this, most of Project Indigo's new experimental features are buttons that generate more deterministic, preset outputs rather than requiring user-written instructions.

The photo critique feature provides what the app labels a "professional" opinion covering framing, lighting, colors, and emotional impact. A companion feature, capture and edit suggestions, recommends how to reshoot or modify an image by adjusting framing, exposure, and removing specific objects in the viewfinder. For example, the app can identify unwanted elements like a hexagonal white object and suggest their removal. The second section guides users on how to use existing photos to improve them via Adobe Lightroom controls.

Object removal has been a staple of photography apps for years, but Project Indigo's implementation improves on the manual process most competing apps require. Rather than forcing users to draw or circle objects on screen—a method that often produces imperfect results—Project Indigo offers toggles for common unwanted elements: people in the background, trash and trash cans, wires and poles, fences, vehicles, and other clutter. Users can also describe custom objects to remove. Testing showed the feature successfully removed a person and the object he was holding from the background of a photo without creating visual artifacts.

The app also includes AI-generated depth of field, which simulates a blurred background on flat photos. Additionally, a style transfer feature allows users to apply artistic effects—watercolor, pen and ink, ink line with color wash, monochromatic, and backlit subject tones—to images. Adobe is currently powering these features with Google's Gemini-based Nano model but remains open to swapping in alternative models, including its own Adobe Firefly. All new AI features are bundled under an "AI playground" tab and remain in testing phase, available only to select users. While these features may never reach a wider audience, their emphasis on teaching photography principles rather than merely generating corrected images sets them apart from purely automated alternatives.

Adobe's Project Indigo represents a deliberate design philosophy shift away from prompt-based generative AI tools. Marc Levoy's background in developing Google's Pixel camera features informs this approach: rather than ask users to craft the perfect text instruction, the app offers preset buttons that produce more predictable outcomes. The photo critique feature—delivering structured feedback on framing, lighting, colors, and emotional impact—occupies a middle ground between fully automated editing and user-driven prompting. By bundling these tools under an "AI playground" tab and limiting access to select testers, Adobe signals that these are experimental rather than production-ready features. The distinction matters: the company is explicitly testing whether cameras can educate photographers rather than simply correct photos, a framing that appeals to skill-building over convenience-driven consumption.