Caira Camera Rolls Back Gen-AI Features

credits: Camera Intelligence

Caira Camera, developed by Camera Intelligence and previously known as the Alice Camera, drew a lot of attention last year when the company announced plans to integrate Google Nano Banana directly into its imaging pipeline. The idea was to let users make major changes to photos using simple prompts, essentially bringing generative editing into the camera itself. The announcement quickly sparked debate among photographers, and the company now says it is dialling those features back. According to CEO Vishal Kumar, the early version of the tool was intentionally experimental and allowed users to type open-ended prompts to change their images.

That open approach turned out to be more confusing than expected. When Camera Intelligence sent early preview units to creators like Tony Northrup, Jimmy Cheng, and Emily Lowrey, the company found that many people simply typed vague or random prompts instead of carefully describing the edits they wanted. Feedback from those creators, YouTube comments, and small workshops with photographers pushed the team to rethink the system. Instead of an open prompt window, the camera now uses 12 curated buttons and templates for specific edits. These include tools for removing clutter from photos as well as stylised “era” looks that mimic the colour styles of different decades, ranging from the 2000s all the way back to the 1920s.

Beyond the generative tools, the camera also includes several computational photography features that do not rely on generative AI at all. Powered by a Qualcomm Snapdragon processor, the system can stack multiple handheld exposures in low light to reduce noise and improve detail, something smartphones have done for years but is still fairly unusual in interchangeable-lens cameras. The camera’s JPEG files also include a hidden HDR gain map, allowing them to display properly on HDR screens. Camera Intelligence is also experimenting with handheld super-resolution, which could boost the camera’s output from 11 megapixels to 44 megapixels by combining several images into a single frame.