Figma Expands Its Agent Platform with Specs and Code Layers at Config India

Figma held its first-ever Config event in India this week — a telling geographic choice, given that the country is the company's largest market outside the United States, with 40% of the top 100 companies listed on the Bombay Stock Exchange using the platform. The central announcement was an aggressive evolution of the AI agent ecosystem on the platform, positioning Figma not just as a design tool, but as a complete creation platform where entire teams — human and agent alike — decide what to build and ship.

Specs: Markdown that lives on the canvas

The most distinctive launch at the event was Figma Specs, a Markdown-based editor that allows teams to define their design language in a structured, executable way. Unlike a static document, a Spec is a multiplayer container that lives on the Figma canvas — it can embed FigJam diagrams, code layers, designs, and references to any canvas node, keeping text and visual elements side by side instead of relying on screenshots or stale links.

What makes Specs particularly relevant to the future of AI agents is that the Figma Agent can read, reply to comments, and act on a Spec the same way it collaborates with a human colleague. Teams can drag a PRD document, ask the agent to expand it with new ideas from a FigJam brainstorm, and the agent operates with direct visual context — not just isolated text.

Luke Zhang, a product manager at Figma, explained in the announcement: "More and more product builders are working in Markdown, as collaborating with agents becomes standard." Specs is Figma's response to that trend — rather than competing with Markdown as a format, the company elevates it to a living container within the platform.

Specs is in open beta for paid plans, rolling out to Professional, Organization, and Enterprise. The core idea is to provide a centralized reference system for agents — whether internal to Figma or third-party — to follow an organization's design and brand rules.

Code Layers: design meets code on canvas

In parallel, Figma opened the open beta for Code Layers — a feature that adds code layers directly on top of designs in the canvas. Initially created in a closed beta in June, Code Layers allows teams to create interactive prototypes with real code backing, whether by converting a design frame, importing a GitHub repo, capturing a website via Chrome extension, or generating from scratch with the Figma Agent.

The feature is particularly interesting for the design-to-code workflow. A designer can create a layout, ask the agent to turn it into code, and iterate on the result — all without leaving the canvas. This significantly reduces friction between visual prototyping and implementation, one of the oldest bottlenecks in the design flow.

Loredana Crisan, chief design officer at Figma, framed the central problem Code Layers seeks to solve in blunt terms: "There are 17-times more lines of code being produced today than in the past. But when you're looking at releases, only 1.3-times more releases. A lot of that code actually is not necessarily built to ship, but really built to discover what the product is. It's built to design."

The quote reveals a fundamental shift in software development mindset — code today is often a form of exploration, not just delivery. Code Layers formalizes that exploration right on the Figma canvas.

Figma MCP: the protocol that connects agents

A transversal component across all these announcements is the Figma MCP (Model Context Protocol) server, now with support for more than 20 external tools, including Open Code, Dia, Grok Bot, and Indian AI startup Sarvam. The MCP allows external AI agents to operate directly on the canvas — creating code layers, editing specs, generating images — without intermediary tools.

This positions Figma as an agent platform, not just an application. Just as Slack became the platform where bots solve team work, Figma wants to become the platform where AI agents perform design work — within the context of the company's visual design and design system.

Payload and Figma Make: integrated dynamic content

Figma also announced the integration of Payload — an open-source headless CMS acquired last year — with Figma Make, the company's tool for generating dynamic websites and managing content directly. Payload saw its downloads increase sixfold in the last 12 months, explaining the strategic decision to deeply integrate it into the Figma ecosystem.

With this integration, teams can connect designs already created in Figma Design directly to content workflows in Figma Weave, enabling visual assets to be used in multi-channel contexts without leaving the platform.

What this means for the future of design

Figma's strategy is clear: transform the platform into the central point where software is planned, designed, and prototyped — with AI agents acting as parallel workforce, not optional tools. Specs provides structural context, Code Layers gives the agent real code execution capability, and MCP allows any external tool to operate on the canvas.

The move is ambitious. Figma is trying to become the operating system of modern design — and AI agents are the execution system running on top. If the platform can maintain quality, performance, and security through this transition, it will not only dominate the design tools market but become critical infrastructure for the entire software creation flow.

The question is: how many more AI agents need to operate on a single canvas before the traditional design interface disappears completely?

Sources: TechCrunch, Figma Blog, The Verge

✓ Independent sources cross-checked and verified before publishing