Adobe brings AI agents and new creative tools to Photoshop, Premiere

Adobe announced a major expansion of artificial intelligence tools across its creative suite during its MAX event in Miami, adding capabilities to Photoshop, Premiere, Illustrator, and InDesign alongside Firefly updates and a new Model Context Protocol connection. According to TechCrunch's report, the company is also opening access for external assistants such as ChatGPT, Claude, and Gemini to interact with desktop applications directly.

This report details the new prototyping software, desktop editing controls, partner model catalog, and local protocol endpoints introduced across Creative Cloud. The breakdown is intended for designers, video editors, and digital teams evaluating how these automated functions fit into professional production environments.

01Firefly Builder prototype canvas and graphics

Adobe launched a new prototyping application called Builder within Firefly, positioning the tool directly in competition with Figma for design mockups. The tool is currently available in a private beta and allows creative teams to assemble layouts to present project concepts quickly. Designers can test layouts and review presentation ideas within the shared creative workspace.

Alongside the prototyping interface, Firefly gained a dedicated system designed to generate on-brand graphs for business and presentation projects. The utility lets teams produce visual data displays that align with established visual standards without manual charting. This addition gives production teams unified controls for structured visual reporting.

Firefly also introduced an image-editing tool that enables creators to modify isolated elements within an image without altering surrounding objects. Users can adjust specific areas while maintaining the background and adjacent compositional details. This targeted editing avoids unintended distortions across complex graphic assets during localized adjustments.

02Proactive video assembly and quick actions

A new video creation environment in Firefly lets users assemble complete sequences by combining video clips, still images, and audio effects. The interface handles multiple media assets to produce cohesive video timelines without requiring manual composition across separate software packages. Editors can arrange core story beats directly inside the Firefly workspace.

Adobe designed its underlying creative agent to operate continuously in an active state rather than waiting solely for manual commands. By remaining proactive and always running, the agent gathers broader context about current project work and stands ready to assist with editing operations. This persistent awareness helps the system suggest contextual actions during production.

The platform also introduced rapid execution shortcuts designed for frequent production tasks across digital channels. Creators can trigger immediate background removal or generate custom video thumbnails with single actions. These streamlined utilities reduce repetitive preparation steps for online video publishers and graphic editors.

03Creative Cloud relighting and object removal

The desktop Creative Cloud applications received targeted feature updates announced at the Miami conference. Photoshop gained a specialized relighting tool alongside dedicated adjustment sliders for prompt-based generative styles. These controls give artists granular authority over artificial light angles and stylistic prompt influence on image layers.

Premiere introduced an updated color grading system developed specifically for professional editors seeking higher precision in post-production. The video editing suite also added an automated object removal tool capable of clearing unwanted visual elements from moving footage. Video specialists can clean complex scenes without frame-by-frame manual paint work.

Lightroom added photographic touch-up tools focused on portrait correction and subject presentation. The software now includes a correction utility to fix blinking eyes in portrait photography alongside an updated blemish removal feature. Photographers can salvage mis-timed portrait frames and eliminate skin imperfections during high-volume editing sessions.

04Partner model catalog and Topaz releases

Adobe continues to expand its broader ecosystem by integrating third-party generative models directly into its creative applications. The platform currently offers external models from Google, Runway, ElevenLabs, Black Forest Labs, OpenAI, Kling, and ByteDance. Creative professionals can access diverse generative models from multiple providers through Adobe software environments.

Topaz Labs, a recent acquisition by Adobe, introduced two specialized image processing models to enhance graphic assets. The company released Wonder 4, a machine learning model built specifically for upscaling resolution and eliminating image noise. The model targets production workflows that require clean asset enlargement without artifact degradation.

Topaz Labs also released ImageHDR, an image transformation model engineered to convert standard dynamic range pictures into high dynamic range formats. The system enhances tonal depth and luminance across standard photographs to meet modern HDR visual specifications. These additions broaden Adobe specialized imaging catalog for restoration and grading tasks.

05Local model context protocol endpoint rollout

Building on existing integrations with ChatGPT and Claude, Adobe is launching a local Model Context Protocol system to connect Creative Cloud tools directly with external chatbots. The implementation exposes desktop application capabilities from Illustrator, Photoshop, After Effects, and Premiere to external language models. This protocol framework lets assistants like Gemini and Claude execute native creative tasks.

Adobe vice president Deepa Subramaniam explained that the company is evaluating how to expose endpoints across core applications. According to Subramaniam, Adobe wants to observe creator habits rather than restrict access: “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.”

The protocol rollout occurs alongside industry debate over standard adoption across competing design ecosystems. Last month at an industry event in India, Figma co-founder Dylan Field remarked that the Model Context Protocol remains imperfect and primarily advantages frontier artificial intelligence laboratories. Adobe plans to continue monitoring developer adoption across its exposed creative endpoints.