Five workflows that show what Figma Weave is actually for: chaining AI nodes on a canvas to blend two references into a style guide, fan out variations across aspect ratios, run eight distortion filters in parallel, generate rotatable 3D models through Rodin 3D V2, and composite stills into rendered video.

Alexia Danton, Designer Advocate at Figma, walks through seven tactics for stretching Make credits further. The most useful ones are the least obvious: use the Edit tool and ​“Go to source” for small visual tweaks instead of prompting, codify repeated instructions into a guidelines.md file so Make doesn’t relearn your conventions every turn, and reach for Gemini Flash on routine iteration while saving Claude Opus for ambiguity and high-fidelity work.

A short onboarding walkthrough for the agent beta. Worth watching for the suggested starter prompts: generating alternative layouts, starting from scratch using your design system, working together on canvas, and getting feedback.

TBPN’s 20-minute interview with Dylan Field on the day of the Design Agent launch.

“We don’t think designers should generate a one-shot screen and call it a day.” That sentence from the announcement is Figma’s vision for the design agent in a nutshell. The framing is explicitly co-pilot, not auto-pilot: the agent runs in front of you on the canvas, riffs to spark an idea, and then hands it back to your mouse and direct manipulation. Pair that with native access to your libraries, components, and tokens, and the bet is clear – the winning AI design tool is the one that already knows your design system, not the one that generates the prettiest screenshot.

Rolling out gradually in beta over the coming weeks. During beta, the agent won’t consume credits. See also the official announcement at The Figma design agent is here.

Brett McMillin shows a concrete loop: an agent reads a coded export flow, finds every state the developer shipped (success, error, loading, edge cases), and generates fourteen designable frames on the canvas using the design system. From there, the designer riffs on three animation directions, the /sync-figma-token skill flags token drift between code and variables, and a generate_figma_design call produces an annotated side-by-side diff.

Emma Webster’s overview of why MCP exists and what it changes. Without context, AI coding tools work from a screenshot — they see the end result, not the decisions that went into it. The Figma MCP server hands agents structured access to components, tokens, and layout decisions instead. Useful as the conceptual baseline before getting into the applied workflows in the lab.

A conversation between Figma’s Design Director of AI Gui Seiz and engineer Alex Kern on how AI inverts the old economics, code used to be expensive and design cheap, now both are cheap and the bottleneck moves to intent. The companion piece to the two MCP posts and the Code to Canvas tutorial elsewhere in this section.

Round up of four AI workflows Figma sees teams adopting: prototyping in code first and pulling it back to the canvas via Codex to Figma, generating dozens of layout variations on the canvas, building a Figma Make prototype before writing the spec, and using Make kits with MCP to carry design system context into the code. The through-line is that the artifact teams align around is shifting from the mockup to the working prototype.

Yuhki Yamashita, Figma’s CPO, lays out the company’s worldview behind the Design Agent, Make, and Weave launches. When generating a working app is cheap, the bottleneck moves upstream: choosing the right direction and shaping it with care. He proposes a ​“go broad and deep at the same time” workflow, where Make spins up parallel prototypes and Weave becomes the room where teams compare, argue, and refine. A tidy thesis for a launch week, and the tools clearly exist to enact it.

Designer Fund surveyed 900+ designers across 60+ countries and conducted 20+ interviews with leaders from Anthropic, Framer, Linear, Notion, Shopify, Sierra, and Stripe. The headline number: half of respondents have shipped AI-generated code to production, and designers are now using double the AI tools they were a year ago. Designers are quietly absorbing PM and engineering work, but hiring loops, performance reviews, and team shapes haven’t caught up.

Figma Make now supports custom skills — markdown files that capture conventions and workflows you use repeatedly, callable from any prompt with a slash command. Pair /build-from-prd with a Notion connector and any PRD becomes a prototype that matches your standards.

Stephen Haney announces QuiverAI’s SVG generation inside Paper – vector output from text or references, aimed at logos, illustrations, and animations. QuiverAI​’s models have been on my radar since Dann Petty joined the team as a Founding Product Designer.

Paul Bakaus has packaged 23 design commands into a single agent skill that teaches Claude Code, Cursor, Codex, and Gemini CLI how to actually design. Type vocabulary, color systems, motion, spatial logic — the foundations most prompts miss. The Live mode that writes accepted variants back to source is where it gets genuinely interesting.

A useful baseline study on how people actually use AI well. The most uncomfortable finding for designers: in conversations that produce artifacts (code, UI, documents), users are less likely to question the model’s reasoning. Polished output suppresses critical evaluation, even though that’s exactly when it’s most needed.

Brian Lovin shows the Notion design team’s Prototype Playground, a single Next.js repo on Vercel where every designer gets a namespaced folder and a small set of shared Notion-flavored components. The interesting parts are not the scaffolding but the slash commands and skills layered on top: /figma runs a three-phase loop with the Figma MCP and Chrome DevTools MCP until the build matches the source (~80% on the first try), and a find-icon skill writes its own TypeScript search script after the team got tired of Claude hallucinating ​“search-icon.” See also the Stripe and Vercel pieces below for the same pattern at other companies, as well as another interview with Brian at Dive Club.

Owen Williams, design manager at Stripe, walks Claire Vo through Protodash, the internal prototyping platform he’s been building for 18 months. The V1 was a bundle of Cursor rules plus an MCP server wired to Sail, Stripe’s design system, so any designer can open the repo and build a page without ever touching React or routing. V2 wraps the whole thing in a browser UI running on internal dev boxes, with embedded LLM chat, click-to-annotate feedback, a design review mode, and fidelity toggles (monospace, grayscale) to signal work-in-progress. The same pattern as the Notion and Vercel pieces in this section: production design system plus a thin internal harness, calibrated to one team’s review culture in a way no off-the-shelf tool can be.

Hannah Hearth runs a tooling Show and Tell with her team at Vercel and writes up the results: Codex + Claude pair programming, Conductor for parallel agent threads, UI Fork for in-browser variant exploration, and Cleanshot’s Pin tool still earning its place.

“Claude Design can read a design system carefully when the prompt is about the system. When the prompt is about a composition that uses the system, it stops respecting the components and just generates lookalikes.” TJ Pitre spent a few hours testing Claude Design against two real design systems and concludes that the tool references your system, but doesn’t consume it. Claude would happily inline HTML tags with style props instead of importing them from your component library.

“The pitch for Claude Design’s workflow is roughly: I have a design system, I want to generate new product surfaces from it, and I want AI to do most of the lift. That workflow exists today. You can pair Figma with an MCP server like our Figma Console MCP, or with Figma’s own MCP server, or with Code Connect, and then point an AI app generator at it. Lovable, v0, Replit, Figma Make, Claude Code working inside your repo. Your Figma file stays the canonical source. Your codebase stays the production surface. AI does the generation in between.That flow is more linear, more honest about where source of truth lives, and it produces output that actually uses your component library, because the AI is operating inside the repo where the components live.”

Grace Li lists 7 common design smells in GPT 5.5: huge typefaces with tight tracking, lack of textures, bento boxes with unrelated icons, pill-with-dot in hero sections, gradient keywords, grid background, and rounded cards with three nested cards. ​“TLDR: if you’re vibe coding with GPT 5.5, the easiest wins are: loosen the tracking on your headlines, delete the status pill, and pick one accent color instead of a gradient. That’ll get you out of the uncanny valley before the next model release closes the gap on its own.”

GPT 5.5 is strong at programming, but not great at visuals: ​“Across 5,000+ preference pairs, GPT‑5.5 ranks #13 in Website Arena, #16 in UI Component Arena, and #19 in 3D Design Arena. It loses to Claude Opus 4.7 and Sonnet 4.6, to Gemini 3.1, to GLM 5.1, and to Kimi K2.6.”