Adobe's Nextcam Team has added a suite of generative AI features called AI Playground to Project Indigo, the company's experimental iPhone camera app available on the Apple App Store. The features were reported by tech outlets on July 20, 2026 (9to5Mac, PetaPixel).

Project Indigo was introduced by Adobe in 2024, about a year before the AI Playground rollout (Engadget). The app serves as a testbed for camera and imaging concepts that may eventually find their way into Adobe's broader product ecosystem. Adobe maintains an official FAQ page for the project on its helpx.adobe.com domain (Adobe Helpx).

The initial version of AI Playground offers four sections: Object Editing, Styles, Photo Guidance, and Custom Edit (Engadget).

Object Editing handles what would traditionally require multiple tools in desktop software like Photoshop. It can remove distractions such as people, cars, or signs from a frame, and it can blur the background to simulate a shallow depth of field — the kind of computational bokeh (the blurred-background effect popularized by smartphone portrait mode) that has been a staple of phone camera pipelines for years, but applied here within a generative editing context rather than at the moment you take the shot.

Styles converts photos to resemble drawings or paintings using presets such as "pen and ink" or "ink line with color wash." The preset names suggest an orientation toward traditional illustration techniques rather than the broader, less controlled stylistic filters common in consumer apps. This puts the feature closer to Adobe's existing stylistic tooling in products like Fresco and Illustrator than to the one-tap filters in social camera apps.

Photo Guidance provides critiques of photos and can offer suggestions before the user takes a shot. This pre-capture advisory capability is the most architecturally interesting of the four. It implies that the app is analyzing a live preview stream, identifying compositional or technical issues, and delivering feedback in real time — a workflow that blurs the line between a camera app and a photography instructor. How the guidance model is trained and what compositional rules it prioritizes are details Adobe has not publicly elaborated on in the available reporting.

Custom Edit allows users to alter images through text prompts, placing it in the same category as natural-language image editing features that have spread across the industry over the past two years. The combination of all four sections within a single mobile camera app, rather than a desktop or web-based photo editor, is the distinguishing structural choice here. Adobe is pushing generative image manipulation closer to the point of capture.

AI Playground was free at launch, but availability was limited to a small fraction of users — a few percent — for only a few weeks. The Nextcam Team stated that if the feature proves popular, it will eventually become a paid feature (Engadget). This gated, time-limited rollout is consistent with Adobe's broader pattern of testing experimental features in controlled cohorts before committing to a pricing or packaging decision.

The broader context here is Adobe's AI monetization strategy. The company has invested heavily in generative image models across its product line, and the question of how those capabilities are priced at the consumer tier — subscription, per-feature, or bundled — remains actively contested across the industry. Project Indigo's AI Playground may serve as one data point in that ongoing calculation.

What makes this deployment worth watching is not any single feature in isolation. Computational photography (the software-driven processing that happens inside your phone's camera) and generative image editing have each matured as separate technical streams. Bringing them together inside a mobile camera app, at the moment of capture rather than in post-processing, is a structural bet on where the editing workflow itself should live. If the guidance and editing tools prove sticky with users, the implication is that the boundary between taking a photo and modifying it continues to dissolve — not in a desktop application, but in the few seconds between preview and shutter.