Design Thursday #101
A weekly recap of everything you need to know about tools, events, guidelines and design in general.
Figma Updates
Make preview on the mobile app
Figma Make previews are now available in the Figma mobile app. You can open a Make on your phone to test it with real gestures, tap through flows, and validate scrolling and interactions the way a real user would.
ChatGPT Images 2.0
OpenAI's latest image model is now available in Figma. You can use it with Make Image and Edit Image across Design, Draw, Slides, Buzz, FigJam, and Figma Weave. It's better at generating infographics, handling multilingual text, and preserving faces across edits compared to the previous model.
Figma Weave workflow templates
Figma Weave templates are now available on the Figma Community. You can browse and duplicate workflows built by the Figma team, from image generation and video creation to scaling brand guidelines into illustration sets. A good starting point if you haven't explored what Weave can do yet.
Explore Weave templates on the Community
Framer Updates
Logo Shaders
Framer added Logo Shaders, a new type of shader built specifically for SVG and PNG logos. Upload your logo and choose from Gradient or Glass effects. The animations respond to the edges of your shape, and effects like Contour, Dispersion, and Bevel add some convincing depth. Great for teaser pages or product launches that need something more polished than a flat logo.
CMS 3.0
The Framer CMS got a complete redesign. The new table view supports inline editing across all cells, so you can update statuses, images, references, and more without opening an overlay. It also adds multi-cell selection, bulk actions, column resizing and reordering, folder support for collections, and a dedicated fields panel. Keyboard navigation has been rebuilt from scratch, and filtering and search are now properly integrated.
Rive Updates
Two editor updates shipped this week. The bigger one adds auto-zoom for the Timeline Graph Editor, file sharing links so you can invite anyone to view a file, drag-and-drop support in the file browser, and extended zoom in the State Machine view. The second update adds keyframe staggering (hold Command + Option on Mac), image fit and alignment for data-bound images in layouts, and the ability to duplicate state machine transitions via right-click. Both updates also include a solid round of bug fixes.
Webflow Updates
Rive animations in GSAP interactions
You can now control Rive animations directly from the Webflow canvas using GSAP interactions. Bind interaction triggers to ViewModel data inside your Rive file, and they'll respond in real time to scrolls, clicks, or any other trigger. Color properties, state machine transitions, and artboard states are all supported, and it all runs on the newer WebGL2 renderer, which means vector feathering for smoother anti-aliased edges.
GSAP interactions in design systems
GSAP interactions can now be built directly into components and shared libraries. You can define motion on main components and variants, include them in shared libraries, and animate individual instances differently for contextual use. It's a big step toward making motion a first-class part of your design system rather than something bolted on per page.
Redesigned code editors
All four of Webflow's code surfaces (Code Block, Code Embed, Page Settings, and Dashboard custom code) now share a single redesigned editor. It includes find and replace, Emmet abbreviation support, inline auto-complete for class names and variables, Prettier formatting, and a VS Code-inspired color scheme. The editors also open in resizable, draggable modals instead of the old fixed panel.
Webflow AEO
Webflow announced AEO, a new product focused on Answer Engine Optimization. It connects measurement, recommendations, and execution in one place. You can track how often your brand shows up in AI-powered search, get prioritized suggestions for what to fix, and have agents ship those changes directly. Currently in private beta for Enterprise customers.
Jitter Displacement shaders
Jitter added a new Displacement shader effect. You can add wave-like distortion to any element by setting amplitude, offset, and repeat values for both the start and end states of an animation. It's useful for fluid transitions, reveal animations, and ambient background motion. You'll find it in the Animate tab under Add Animation.
Lottie Creator MCP server
Lottie Creator now has an MCP server, so you can connect Claude, Cursor, or any MCP-compatible client and control the editor through prompts. That means generating scenes and keyframes, editing colors and easing, and even batch-theming animations for dark mode or brand variants, all without leaving Creator. They also added a brand-check feature where you can pass in a URL and your AI will flag anything that doesn't match.
Recraft is Now in Figma, Framer, and Chrome
Recraft launched plugins for Figma, Framer, and Chrome, all powered by the V4 model. In Figma you can generate assets, real SVG vectors, and icons directly on the canvas. In Framer you can create web-ready visuals and vectorize or upscale images without leaving your project. The Chrome extension lets you generate images from any webpage and drop them directly into Google Docs, Slides, or Gmail. You will get 30 free credits daily across all three.
Penpot MCP Demo
The Penpot team published a demo showing an AI agent connected via MCP helping with a full mobile UI design session. Using just a few prompts, the agent generates color palettes, typography scales, icons, and other elements directly inside the shared canvas. Worth watching if you want to see what co-designing with an AI agent looks like in practice.
Codex Gets a Major Update
OpenAI shipped a big update to Codex that goes well beyond coding. It can now use apps on your Mac with its own cursor, work in multiple agents in parallel, generate images, browse the web in a built-in in-app browser, and remember preferences from previous sessions. New developer-focused features include GitHub PR review support, multiple terminal tabs, SSH access to remote devboxes, and over 90 new plugins for tools like Jira, GitLab, CircleCI, and more. It's starting to look less like a coding tool and more like a full AI workstation.
Brian Lovin on leveling up as a designer with AI
Dive Club published a conversation with Brian Lovin, design engineer at Notion, on how he's evolving his design process with AI. It covers everything from prototyping workflows and prompting strategies to his current toolset and a new side project he's working on. Worth watching if you want a practical look at how someone at the top of the field is actually integrating these tools day to day.
Introducing Claude Design
Anthropic launched Claude Design, a new product that lets you create prototypes, slides, one-pagers, and other visual work through conversation with Claude. It's powered by Claude Opus 4.7 and supports inline comments, direct edits, and custom sliders for fine-tuning. Teams can connect their design system so all output stays on-brand, and finished designs can be exported to Canva, PPTX, HTML, or handed off directly to Claude Code.
Why do AI models hallucinate?
Anthropic published a video explaining what researchers mean by hallucination, why it happens, and how to spot it in your own conversations. A good watch if you work with AI tools daily and want a clearer mental model of what's actually going on under the hood.
UX Tools Survey: Three Groups of Designers
UX Tools published results from their 2025 survey of 1,478 designers, and the main finding is that the design field has split into three distinct groups with very different relationships to AI. 43% are already vibe coding, 37% haven't tried it at all, and the 18.5% in between are the ones still deciding. 59% of designers are now building their own tools, and there's a satisfaction gap between those who've adopted AI workflows and those who haven't. The full breakdown is worth reading.