Adobe unveils AI tools across creative apps, plans MCP private beta for 2027

The MAX announcements extend Adobe's push to embed generative AI into Creative Cloud, building on existing ChatGPT and Claude integrations and Firefly tools, and adding technology from Adobe's Topaz Labs acquisition.

Among the new features in Photoshop is the Relight tool, which Adobe says lets users non-destructively add, reposition and resize light sources in a composition, alongside Generative Restyle for generating variations of an image's style with adjustable sliders. Beta features include Mockups for compositing designs onto objects, Camera Angles for exploring new viewpoints, Generative Recompose for changing aspect ratio, Convert to Text Layers, Timelapse and a tool that removes ambient effects such as lens flare without removing objects. Adobe also introduced two image enhancement models from its Topaz Labs acquisition: Wonder 4, combining upscaling, denoising, deblurring and sharpening in one operation, and Image HDR.

In Premiere, Adobe is adding a Generative Media Tool that brings video, sound effects, soundscapes and music generation into the timeline, drawing on Adobe Firefly and partner models from Google, Kling, Runway and Luma, and able to generate missing b-roll or background sound from the context of an edit. The beta adds Generative Remove for deleting unwanted objects such as boom mics and logos, while a new Color Mode color-grading tool is available starting today. Premiere Mobile gains licensed music tracks from UMG and WMG for non-commercial use, with Quick Cuts and an AI Assistant expected soon. Lightroom adds Glow, Blemish Removal, Open Eyes in beta and expanded camera tethering.

On the Firefly side, Adobe is adding Firefly Video, which organizes scenes, assembles clips, adds captions and generates transitions and b-roll, and Firefly Design, a single canvas for illustration, image editing, animation and 3D. Firefly Builder, in private beta, prototypes websites and apps from natural language prompts. Firefly's AI assistant gains Plan Mode, for reviewing and approving an agent's proposed actions, and a Model Selector, while Adobe previewed always-on agents for background tasks such as organizing assets and preparing content for different platforms. AI assistants are now in beta in Illustrator, Premiere, InDesign, After Effects, Frame.io and the Photoshop beta. Illustrator also receives new brushes, gradients and blend controls, and InDesign adds a co-editing beta, a Flex Layout tool and document-checking features.

Longer term, Adobe is preparing a Model Context Protocol integration so its apps can work with AI assistants such as ChatGPT, Claude and Gemini. The company said it will run locally on a user's computer, letting an agent work directly in native project files while layers, vectors, bins and sequences stay editable, and that it will enter private beta in early 2027. Adobe vice president Deepa Subramaniam said the company is still deciding what to expose. “We are still figuring out where and when and how to draw the line. What we're putting out there is a number of endpoints and capabilities across Illustrator, Photoshop, After Effects and Premiere. We're just going to sort-of see how people use it, what they're trying to do, and keep showing more. We don't have any need or desire to sort of limit what the MCP endpoint can expose. What we really want to do is to support that creative workflow, how you choose to do it,” she said. Figma co-founder Dylan Field said last month at an event in India that MCP is not a perfect standard and benefits frontier labs more.