Adobe announced in July 2026 that it is adding a set of AI-powered features to Project Indigo, its experimental iOS camera app. The new capabilities include LLM-based photo critiques, editing suggestions, advanced object removal, depth-of-field generation, and style transfer.
Project Indigo was launched last year and previously offered pro camera controls, multi-frame super-resolution, and various capture modes. The new features are grouped under an “AI playground” tab and are currently available only to select users, with no guarantee of a wider rollout.
Marc Levoy, who leads the project and previously worked on camera development for Google’s Pixel phones, said that prompt-based AI editing can be difficult because finding the right prompt to achieve a specific result is often unreliable. Most of the new features are designed as buttons that produce more consistent, deterministic outputs rather than relying on open-ended text prompts.
The photo critique feature offers feedback on framing, lighting, colors, and emotional impact. A companion feature provides capture and edit suggestions, advising users on how to reshoot or adjust a photo, and also recommends specific edits that can be made using Adobe Lightroom controls.
The object removal tool differs from similar features in apps like Apple Photos, Google Photos, and Adobe Photoshop, which typically require users to manually select objects by drawing on the screen. Project Indigo instead offers toggles for common categories — such as background people, trash, wires, poles, fences, and vehicles — along with an option to describe a custom object for removal.
The style transfer feature can apply visual treatments including watercolor, pen and ink, monochromatic, and backlit subject. The app also includes a text-based editing prompt for edits not covered by preset tools.
Adobe is currently using Google’s Gemini-based Nano Banana model to power these features, but has indicated it is open to incorporating other models, including its own Adobe Firefly.
Source: TechCrunch