Figma adds Specs and public beta code layers for AI-assisted design
Figma is giving teams more ways to define design rules for agents and test code-backed prototypes inside its platform.

Figma is adding new ways for teams to tell AI agents how design work should be done inside its platform, according to TechCrunch. The company announced a Markdown-based editor, broader access to code layers, new MCP tool support and a Payload CMS integration at its first Config event in India.
For designers, product managers and small software teams, the news points to a clearer direction for Figma: less focus on one-off AI prompts, and more on shared rules, reusable context and prototypes that can move closer to working software.
Specs gives agents a source of truth
The headline addition is Figma Specs, a Markdown editor designed to help companies describe their design language in a structured but simple format. Figma says the document can be used by both human teams and agents as a reference when creating designs inside the platform.
Specs can be its own file in Figma or live alongside work on the canvas. Beyond plain text, users can point to other Figma assets, including canvas elements, FigJam diagrams, code layers and designs.
That matters because agents need context to behave consistently. A brand system is rarely just a logo, font and color palette; it includes component usage, interaction patterns, accessibility expectations, naming conventions and internal rules that teams learn over time. Figma is trying to make more of that context available in one place.
The company also said teams can bring in design libraries, code libraries and guidelines to manage design system context. Agents can refer to that centralized material so their output stays closer to the rules a team has already set.
The move fits a wider pattern in which software companies are building guardrails around agent behavior, not just adding more automation. We have seen similar attention to defining agent intent and boundaries in areas such as commerce, including Amex’s intent framework for AI shopping agents and Shopify’s work with AI shopping agents.
Code layers move into public beta
Figma is also expanding access to code layers, a feature that adds code to designs on the canvas to support more interactive prototypes. The company first announced code layers in private beta in June, and is now making the feature available in public beta so more users can test it.
The practical benefit is that a prototype can become more than a visual mockup. According to TechCrunch, users can ask Figma’s agent to turn a design into code or modify part of the code as they iterate.
For freelancers and small teams, that could reduce some of the back-and-forth between design exploration and early technical testing. It does not mean every design becomes production-ready software. But it can help teams test interactions, communicate intent to developers and spot issues earlier than a static mockup would allow.
Figma’s chief design officer Loredana Crisan told TechCrunch that the company sees Figma as where software teams decide what to build and ship. She also pointed to a gap between the growth in code being produced and the smaller increase in releases, arguing that much of the code being created is used to discover what a product should be, rather than to ship directly.
More tools connect through MCP and Payload
Figma said it is adding support for more than 20 new tools through its MCP, including Open Code, Dia, Grok Bot and Sarvam, an Indian AI startup. The company did not list every new tool in the TechCrunch report, but the direction is clear: Figma wants outside agents and tools to work with its design context instead of operating separately.
That could be important for teams already experimenting with agentic workflows across design, coding and content. If agents can read shared design rules and interact with project assets through defined connections, they may produce work that is easier to review and less likely to drift from the system a team maintains.
The company also announced an integration between Payload and Figma Make. Figma acquired the open-source headless CMS last year. Now, Payload is being tied into Figma Make so users can generate dynamic websites and manage content directly, according to the company.
Figma also said Payload downloads have grown sixfold over the past 12 months. For creators and small businesses, the most relevant point is not the download figure itself, but the combination of design, content management and website generation in one workflow.
Why India was part of the story
The announcements came at Figma’s first Config event in India, where the company has been expanding its presence. Figma opened its first India office late last year, with sales, partnership, go-to-market and product support functions.
The company said India is its largest user market outside the United States. It also said the platform works with 40% of India’s top 100 companies listed on the Bombay Stock Exchange.
For Emonarc readers, the main takeaway is that Figma is treating AI agents as collaborators that need documentation, constraints and access to shared systems. If your team is already using Figma as a design hub, Specs could become the place where brand and product rules are written for both people and machines. If you are testing AI-assisted design-to-code workflows, the public beta of code layers is the feature to watch next.
Frequently asked questions
What is Figma Specs?
Figma Specs is a Markdown-based editor that lets teams define their design language and provide a reference for designers and agents working in Figma.
What changed with Figma code layers?
Figma is making code layers available in public beta after announcing the feature in private beta in June.
What is Figma doing with Payload?
Figma is integrating Payload with Figma Make so users can generate dynamic websites and manage content directly, according to the company.



