Artificial Intelligence

Adobe’s Indigo Camera Uses AI to Critique and Reshoot Photos

Project Indigo adds LLM-powered feedback, object removal, and style transfer in a testing-phase 'AI playground.'

Adobe is adding new AI-powered features to Project Indigo, its experimental iOS camera app launched last year, and the headline addition is a feature that critiques your photos. Using large language models, the app now offers a “professional” opinion on framing, lighting, colors, and emotional impact, then suggests how to improve the shot. TechCrunch reports the tools are bundled under an “AI playground” tab still in testing.

The app previously offered pro controls, multi-frame super-resolution, and different capture modes. The new layer is different: instead of prompt-based editing, most of the experimental features are buttons that generate more deterministic outputs. Marc Levoy, the person heading the project and the engineer behind Pixel’s camera chops, had criticized prompt-based editing as fiddly—finding the perfect prompt to tweak a photo can be a chore.

There are two features in the critique category. First is the photo critique itself, with feedback on composition and mood. Second is capture-and-edit suggestions that tell you how to reshoot—for instance, the app told the reviewer to remove a hexagonal white object from a frame. A separate section suggests how to improve an existing photo using Adobe Lightroom controls.

Separately, they sit in an “AI playground” tab available only to select testers for now.

What Adobe’s Indigo AI can actually do

Beyond critique, Project Indigo adds advanced object removal with toggles for things to strip from an image: people in the background, trash cans, wires and poles, fences, vehicles, and other clutter—plus a field to describe a custom object. The reviewer noted the app removed a friend and the object he was holding from the background without strange artifacts, a step up from the circle-and-select removal older tools require.

It also generates depth of field to simulate a blurred background, and experiments with style transfer that turns a picture into tones like watercolor, pen and ink, or monochromatic backlit subjects. Levoy’s team is using Google’s Gemini-based Nano Banana for these features but says it is open to swapping in other models, including Adobe’s own Firefly.

The tension Levoy identified is real: the same descriptive, button-driven approach that makes Indigo’s suggestions useful also opens a road to slop if users lean on presets instead of judgment. Adobe is threading that needle by keeping the features in a testing phase with select users.

Why a camera app, and why now

Project Indigo is a research showcase as much as a product, a place for Adobe to demonstrate nuanced, photography-literate AI rather than one-click filters. It sits alongside the company’s larger generative push tracked in coverage of AI video and world models.

The framing also reflects a wider debate about an AI-saturated media landscape, where the differentiator becomes taste and intent rather than raw generation—the same worry behind concern over an AI-built web.

These features might never reach a wide audience, but they are a useful signal: Adobe is exploring AI that teaches photography instead of just manufacturing more of it.

FAQ

What is Adobe Project Indigo?

Project Indigo is Adobe’s experimental iOS camera app, launched in 2025, that offers pro capture controls and now adds LLM-powered photo critique and editing suggestions.

Which AI model powers the new features?

Adobe is using Google’s Gemini-based Nano Banana for the features but says it is open to swapping in other models, including its own Firefly.

Are the features available to everyone?

No. The tools sit in an “AI playground” tab that is still in a testing phase, with only select users getting access for now.