Adobe adds an AI Playground to its experimental Project Indigo camera app

Adobe has added a bundle of generative-AI features to Project Indigo, its experimental iOS camera app, grouping them under a new “AI Playground” tab, according to reporting published July 20, 2026. The standout is a photo-critique tool that uses large language models to assess a shot.

Point it at an image and the critique feature grades framing, lighting, color and even emotional impact, then suggests changes or a reshoot, turning the app into something like an automated photography tutor. The update also brings advanced object removal that can toggle off people, wires, poles, trash and vehicles without manual selection, plus AI-generated background blur and style transfers into looks like watercolor, pen-and-ink and monochrome.

Project Indigo, led by imaging researcher Marc Levoy, is Adobe’s testbed for computational photography, and the AI Playground pushes it from capture into editing. The features run on a cloud-based model, reported to be Google’s Gemini-based system nicknamed “Nano Banana,” so they need an internet connection and cap edited photos at 2K pixels, a real limit for anyone shooting for print or large displays.

For now this is a small test, not a launch. Access is limited to a slice of Project Indigo’s users for a few-week trial and is free during testing; Adobe has not set broader-rollout timing or pricing. That the editing runs on a third-party model rather than Adobe’s own Firefly system is notable given how heavily Adobe markets Firefly as commercially safe AI, and the company has not said whether that would hold in a wider release. Whether the AI Playground graduates into Adobe’s shipping apps, or stays a research demo, is what to watch as the trial runs out.