Adobe has added a suite of AI-powered features to Project Indigo, its experimental iOS camera app, including a real-time photo critique tool that evaluates framing, lighting, colors, and emotional impact (TechCrunch).

Project Indigo launched in 2025 under the direction of Marc Levoy, who heads the project at Adobe. The app functions as a computational photography testbed on iOS, and the new features, reported by TechCrunch on July 20, 2026, push it deeper into territory that blends capture-time guidance with post-capture generative editing.

The critique feature goes beyond metadata-based analysis. It provides AI-generated opinions on composition, color choices, and what Adobe describes as emotional impact. Alongside the critique, the app offers AI-powered capture and edit suggestions that guide users on adjusting framing, exposure, and objects visible in the viewfinder. The distinction between critique and suggestion matters: one evaluates what the user has already done, the other coaches them in real time before the shutter fires.

Object removal is the most granular of the new additions. Project Indigo provides preset toggles for common photographic distractions: people in the background, trash and trash cans, wires and poles, fences, vehicles, and a catch-all category for other clutter. Users can also describe a custom object they want removed from the image, pointing generative fill at a specific target rather than relying on category-based detection.

A depth-of-field generation feature simulates blurred backgrounds, the kind of shallow depth-of-field effect typically associated with larger sensors and fast glass. The computational approach mirrors what smartphone makers have been doing with portrait modes for years, but placing it inside Adobe's camera app means the effect can be applied with more sophisticated segmentation and at a point in the pipeline where Adobe controls the full image stack.

Style transfer rounds out the feature set. Available styles include watercolor, pen and ink, ink line with color wash, monochromatic, and backlit subject. These are applied as generative re-renderings of the captured image rather than as simple filter overlays.

The feature set taken together positions Project Indigo as something other than a conventional camera app. The critique and suggestion features embed an AI advisor into the capture workflow. Object removal and depth-of-field generation bring desktop-grade editing tools to the moment of capture. Style transfer extends the pipeline into generative reinterpretation.

Worth flagging: the critique feature in particular introduces a qualitative judgment layer into a camera app. Evaluating "emotional impact" is a subjective, aesthetic assessment, not a technical one, and embedding it into a capture tool raises questions about what baseline the AI uses to render those judgments and whether it steers users toward a homogenized visual style. Adobe has not detailed the training data or evaluation criteria behind the critique model.

The object removal presets also reflect a particular set of assumptions about what constitutes visual clutter. Trash cans and utility poles are standard urban scenery for street photographers; a preset that removes them by default nudges users toward a sanitized visual aesthetic. The ability to toggle these categories individually gives the user control, but the taxonomy itself encodes choices about what belongs in a frame.

Adobe's decision to put these capabilities into an experimental app rather than a flagship product is consistent with how the company has historically tested computational photography features before deciding whether they graduate to Lightroom or Photoshop. The app's existence under Marc Levoy's direction signals serious investment, and the feature set as a whole suggests Adobe is exploring how much of the traditional post-processing workflow can be collapsed into the capture moment.

For developers and photographers watching the convergence of generative AI and computational photography, Project Indigo's new capabilities are an early read on where Adobe sees the camera app heading: not just a capture tool, but an AI-guided pipeline from viewfinder to finished image.