What happened
TechCrunch reports that Figma introduced Figma Specs at its first Config event in India. The Markdown-based editor lets teams document design language and working rules, with links to canvas elements, FigJam diagrams, code layers and completed designs. Figma also opened Code Layers to public beta, expanded MCP support and announced a Payload integration with Figma Make.
TechCrunch reports that Figma Specs can exist as a separate file or sit directly on a project canvas. Teams can centrally add design libraries, code libraries and guidelines to the design system. According to Figma, AI agents will be able to use that context to follow established brand rules more accurately when generating designs. The source provides no independent testing of that claim.
TechCrunch also reports that Code Layers, previously tested in a closed beta in June, is now in public beta. The places code on the design canvas for more interactive prototypes, and users can ask an agent to turn a design into code or modify code during refinement. The report does not specify public-beta eligibility, supported plans or pricing.
Figma announced support for more than 20 additional tools through the Model Context Protocol, including Open Code, Dia, Grok Bot and Sarvam, according to the report. Figma is also integrating Payload, an open-source headless CMS it acquired last year, with Figma Make for dynamic websites and content management. Figma says Payload downloads rose sixfold over the previous 12 months; that figure is not independently confirmed in the source.
Why it matters
Figma Specs addresses a practical problem in AI-assisted design: agents need structured, shared context to apply a company’s visual and technical conventions consistently. If the works as described, teams could spend less time correcting generated designs and prototypes. The report does not independently verify the claimed accuracy benefits, and it does not establish which users can access Specs or what either feature costs.
The update positions Figma’s design system as operational context for AI agents rather than only documentation for human designers. That distinction matters because generated interfaces can drift from brand, accessibility or implementation rules when an agent lacks local project knowledge. A shared reference could make agent output more predictable, but the source does not show measured reductions in errors, review time or rework.
TechCrunch quotes Figma Chief Design Officer Loredana Crisan saying that 17 times more lines of code are being created than before, while releases have increased 1.3 times. Crisan said much of that code is exploratory rather than intended for release. If accurate, Code Layers may help teams use that increased coding activity for rapid product exploration, but the source gives no methodology or independent validation for the comparison.
Interactive Mechanism: How It Actually Works
Explore the underlying technology behind this development interactively.
crm_get_transaction(id='4092').An agent must create a draft calendar event for Tuesday at 2 p.m. Which evidence would establish the requested result?
What to watch next
The key unknowns are access, pricing and performance. The source does not say whether Specs is generally available, which Figma plans include it, or which agents can use its context. Readers should also watch whether Code Layers moves beyond public beta, how broadly the new MCP connections work, and whether Figma provides evidence for its reported changes in code creation and releases.
Figma has not, in the supplied report, disclosed when Specs will be broadly available, whether it is free or paid, or whether access varies by plan. It also does not identify the full list of supported agents or explain how permissions and sensitive design-system information are managed.
The public beta should clarify whether Code Layers produces maintainable production code or is mainly a prototyping aid. Follow-up reporting should distinguish agent-generated code used for exploration from code shipped to customers.
The breadth of the MCP rollout and the Payload integration will be important practical tests. The source reports announcements and company metrics, but provides no independent performance results, customer examples or security assessment.