TL;DR: Adobe used its MAX conference in Miami to push AI deeper into Photoshop, Premiere, Lightroom and Firefly, and opened a local MCP server that lets outside chatbots like ChatGPT, Claude and Gemini reach into Creative Cloud directly. The headline new piece is Firefly Builder, a private-beta prototyping tool that puts Adobe on Figma’s turf.
Adobe doesn’t usually announce a dozen things at once and expect anyone to remember all of them a week later. This year it tried anyway.
The centerpiece is Builder, a new tool inside Firefly that lets someone sketch out an idea and turn it into a working prototype, the kind of job that used to live entirely inside Figma. It’s in private beta for now, which tells you Adobe isn’t ready to claim victory yet, just a foothold.
Around Builder sits a pile of smaller, more immediately usable features. Photoshop gets a relighting tool and prompt-driven sliders for generative styles, so instead of hunting through menus you describe the look you want and drag a bar until it lands. Premiere picks up a color grading system built for people who already know what they’re doing, plus a way to erase objects from video without the usual rotoscoping grind. Lightroom quietly adds fixes for blinking eyes and blemishes, the unglamorous stuff that eats up an editor’s afternoon.
Firefly itself picked up three features worth knowing about: a chart generator that actually respects your brand guidelines instead of spitting out generic bars and pies, a partial-edit tool that changes one object in an image without disturbing everything around it, and an AI video assembly feature that stitches clips, stills and sound effects into something coherent. None of that is as flashy as Builder, but it’s the kind of thing a working creative team will actually open on a Tuesday.
Then there’s the Firefly agent, which Adobe describes as proactive and always running in the background of your project. It watches what you’re doing and offers to act, removing a background or spinning up a video thumbnail without being asked twice.
The more consequential move, at least for anyone who thinks about where software is heading, is the new MCP server. MCP, short for Model Context Protocol, is the connective tissue that lets one AI system reach into another application and actually do things rather than just describe them. Adobe’s version opens up endpoints across Illustrator, Photoshop, After Effects and Premiere, so a chatbot like Claude or ChatGPT can trigger real actions inside those apps instead of just generating a mockup of what the action might look like.
“We don’t have any need or desire to sort of limit what the MCP endpoint can expose,” Adobe vice president Deepa Subramaniam told TechCrunch, adding that the company is still deciding which capabilities to open up based on user feedback.
That’s a notably loose stance for a company that has spent decades protecting its file formats and workflows. It also puts Adobe in an odd spot relative to Figma, whose co-founder Dylan Field said last month that MCP “is not a perfect standard” and tends to benefit the big AI labs more than the companies building on top of it. Adobe building its own on-ramp for those same labs, while simultaneously launching a Figma competitor, is the kind of move that only makes sense if you assume the AI assistants are the new browser, and you’d rather be the page they load than get left out entirely.
Adobe isn’t building every model itself, either. It already licenses generation tech from Google, Runway, ElevenLabs, Black Forest Labs, OpenAI, Kling and ByteDance, treating Creative Cloud more like a marketplace than a single pipeline. Topaz Labs, which Adobe bought earlier this year, is shipping two new models under the MAX banner: Wonder 4 for upscaling and denoising, and Image HDR for converting flat images into HDR versions.It’s a lot to absorb from one keynote, and Adobe clearly knows it, since most of it ships as a private beta rather than a feature you can click on today.
The bigger pattern is one we keep seeing across the industry this week: big platforms racing to let outside AI agents act inside their products rather than just answer questions about them. Anthropic made the same bet days ago with its own collaborative Meetings, Whiteboards and Sheets tools for teams, and OpenAI has been quietly stitching travel and booking functionality into ChatGPT through a small acquisition of its own. Everyone wants to be the surface the agent acts through, not the app the agent merely describes, which is the same tension we saw play out this week in a hackathon testing how far AI coding agents will bend once another agent offers to help.
There’s a business logic underneath all the feature announcements worth spelling out. Adobe’s Creative Cloud subscriber base is enormous, but growth in a mature market mostly comes from either raising prices on people already paying or pulling in a new category of user who wasn’t a customer before. Builder is squarely aimed at that second group, the product managers, founders and marketers who currently reach for Figma or Canva because Photoshop and Illustrator always felt like tools built for people who already know what they’re doing. If Builder actually lowers that bar, Adobe gets a shot at customers it’s never successfully converted before, without needing to discount anything for its existing base.
Canva, for its part, has spent the last few years doing the opposite move, climbing upmarket from simple templates toward more serious design work, which makes this MAX announcement feel less like Adobe innovating in a vacuum and more like two companies converging on the same middle ground from opposite directions. Whoever gets there first with a product good enough to actually replace the other tool, rather than just compete alongside it, wins a meaningfully bigger slice of a market that used to be cleanly segmented by skill level.
The bottom line
Bottom Line: Builder is the feature that will get the headlines, but the MCP server is the one that changes how Creative Cloud behaves long term. Adobe just made its apps something an outside AI can operate, not just generate content for, and that’s a bigger bet than any single new tool.
The Teck Hustlers Editorial Team covers technology, business, finance and gaming news for readers who want the story behind the headline. Every article is researched, written and edited by our in-house team, checked against company announcements, filings and reputable reporting, and updated when new facts emerge. We focus on what each development means for founders, professionals, investors and everyday users. Have a tip or a correction? Reach us through our contact page.
