Adobe Expands Creative Cloud AI and Launches MCP

Adobe expanded its artificial-intelligence strategy at its MAX event in Miami, adding new generative and editing tools across Creative Cloud while also introducing a Model Context Protocol server to extend its capabilities to external chatbots such as ChatGPT, Claude and Gemini. The company is broadening AI features in flagship applications including Photoshop, Premiere, Illustrator, InDesign and Lightroom, alongside its Firefly platform, which is receiving updates tied to a more proactive, always-on creative AI agent. Adobe aims to make the agent more context-aware so it can anticipate next steps in creative workflows and simplify actions such as background removal or video thumbnail creation. In Firefly, Adobe introduced a private-beta prototyping tool called Builder, signaling a move toward more interactive design creation. The platform also gained tools for generating on-brand graphics, selectively editing portions of an image without altering unrelated objects, and assembling video from clips, images and audio elements. These additions reflect Adobe’s effort to position Firefly as a broader generative workspace rather than a standalone AI utility. Across its core apps, Adobe added new AI-powered editing capabilities. Photoshop now includes relighting features and prompt-based style sliders, enabling users to adjust visual tone and appearance through generative controls. Premiere gained a professional-oriented color grading system and video object removal, while Lightroom introduced corrections for blinking eyes and blemish removal. Adobe continues to support third-party AI models from providers including Google, Runway, ElevenLabs, Black Forest Labs, OpenAI, Kling and ByteDance. Its recent acquisition Topaz Labs also released two image models: Wonder 4 for upscaling and denoising, and ImageHDR for converting standard images into high dynamic range output. A key part of the announcement is Adobe’s new local MCP integration, designed to expose Creative Cloud functionality to external AI assistants and agents. According to Adobe Vice President Deepa Subramaniam, the company plans to begin with endpoints across Illustrator, Photoshop, After Effects and Premiere, then refine the offering based on how users adopt it. The initiative is intended to support user-directed creative workflows rather than prescribe a single interface. The MCP rollout also comes amid broader industry debate. Adobe rival Figma has argued that MCP is not a perfect standard and may benefit frontier model developers. Adobe, however, appears to view the protocol as a way to keep its creative software embedded within emerging AI-driven workflows.