Adobe’s Project Indigo adds AI-powered photo critique and editing suggestions in experimental iOS app
The update introduces LLM-based critiques, capture and edit suggestions, advanced object removal, and style transfer features, all accessible via a new AI playground tab.
1 source · cross-referenced
- Adobe’s experimental iOS camera app Project Indigo now includes AI-powered photo critique and editing suggestions using large language models.
Adobe’s experimental iOS camera app Project Indigo, launched in 2025, has added AI-powered features that use large language models to critique photos and provide editing suggestions. The app previously offered pro controls, multi-frame super-resolution, and different capture modes.
The new critique feature provides a "professional" opinion on framing, lighting, colors, and emotional impact, while the capture and edit suggestions offer real-time tips for reshooting, such as adjusting framing, exposure, or removing objects from the viewfinder. Adobe’s Marc Levoy, who leads the project, noted that most generative AI tools rely on prompt-based editing, which can be cumbersome, so Project Indigo’s features are designed as deterministic buttons to generate more consistent outputs.
The update also introduces advanced object removal with toggles for predefined categories such as people in the background, trash cans, wires, fences, vehicles, and other clutter, as well as a custom object removal option. Adobe claims the feature removes objects without creating artifacts, citing an example where it removed a person and an object they were holding from the background.
Additional features include AI-generated depth of field for simulating blurred backgrounds and style transfer options that transform photos into artistic styles like watercolor, pen and ink, monochromatic, and backlit subject. Adobe acknowledges that style transfer is not a novel concept but notes that modern models produce more refined results.
The company is using Google’s Gemini-based Nano Banana model for these features but has indicated it may swap in other models, including its own Adobe Firefly. These experimental features are accessible via a new AI playground tab and are currently available only to select users as part of a testing phase.
- Jul 20, 2026 · Hugging Face
NVIDIA releases Cosmos 3 Edge, a 4-billion-parameter open world model for robotics and vision AI on edge devices
Trust78 - Jul 19, 2026 · TechCrunch — AI
Nonprofit Current AI launches open public AI infrastructure projects with $3.2M in grants
Trust76 - Jul 19, 2026 · AWS — Machine Learning Blog
Smartsheet deploys remote Model Context Protocol server on AWS to connect AI agents to enterprise data
Trust79