Adobe's new AI tools can erase objects from videos, turn lamps on magically

Adobe previews new AI "Sneaks" like Project Frame Forward, Project Light Touch, and Project Clean Take that simplify photo, video and audio editing with smart, intuitive tools

At its Max conference, Adobe showcased a range of experimental AI tools, known as “Sneaks,” that are said to simplify creative work across photos, videos, and audio. According to Adobe, these early-stage projects can instantly apply edits made to a single video frame across the whole clip. Meanwhile, it can also adjust lighting in photos and fix mispronounced words in audio.

 

The sneaks include Project Frame Forward, Project Light Touch, and Project Clean Take, which aim to make editing more intuitive, helping creators save time and effort while achieving professional results. These tools are not yet available to the public, but many of Adobe’s current Creative Cloud features like Photoshop’s “Harmonize” and “Distraction Removal” began as sneaks, suggesting that some of these AI tools could become part of Adobe’s apps soon.

 

Adobe’s Sneaks: What’s new

Project Frame Forward: This is designed for video editors who want to add or remove people or objects from footage without using complicated masking tools. Normally, editors have to manually trace objects frame by frame, but this tool automates the process. Users can delete a person from a scene in one tap, fill in the background naturally, or even insert new objects using text prompts like “add a dog” or “place a table.” The AI keeps these edits consistent across the entire video, similar to Photoshop’s “Content-Aware Fill.” 

 

Project Light Touch: It focuses on lighting in photos. It uses AI to change how light behaves, adjusting its direction, brightness and colour. Users can make it look like lamps are turned on when they were not, drag the light across a room, or even shift a scene from day to night. It can also create glowing or coloured light effects, giving editors full creative control over illumination.

 

Project Clean Take: It brings AI editing to audio. It can fix mispronunciations, change how a voice sounds, for example, making it more exciting or calm, and even replacing certain words without re-recording. The tool can also separate background sounds, so editors can reduce noise or isolate voices for better clarity. 

 

Other experiments

 

Adobe also previewed other experiments, such as Project Surface Swap, which instantly changes materials or textures in images and Project Turn Style, which lets users rotate objects in 3D. These are:

• Project Surface Swap: This tool uses AI to recognise and replace materials in a photo while keeping lighting and perspective natural. Users can change how a surface looks — for example, swap a sofa’s fabric or turn a wooden floor into marble.
• Project Trace Erase: It makes removing unwanted objects easier and more realistic. Instead of just erasing something, it also removes its shadows, reflections, and background distortions, with almost no manual editing needed.
• Project New Depths: This project brings 3D control to photo editing. It lets you adjust colours, shapes, and composition in three-dimensional space. The company said that it is almost like sculpting a picture.
• Project Scene It: It helps artists design detailed 3D scenes while keeping control over each object’s style. It allows tagging of individual objects with reference images so their look stays consistent even as they move around in 3D space with the help of AI. Adobe said that it is a creative tool for building lifelike worlds with precision and artistic freedom.
• Project Motion Map: This tool can automatically animate still images like vector graphics. Adobe said that by studying how the image is structured, it adds movement that feels natural and expressive without needing complex keyframes or manual animation.
• Project Sound Stager: Sound Stager uses AI to design sound based on a video’s visuals and mood. It can automatically build layered sound effects that match the pacing and emotion of a scene. Users can also chat with an AI “sound designer” to fine-tune the final audio mix.