Design software company Figma has unveiled a new set of AI-powered features for its creative platform and added 20 partners to its Model Context Protocol (MCP) ecosystem, including Indian AI startup Sarvam. The announcements, made at Config India in Bengaluru on October 15, signal Figma's push to bring designers, developers and AI agents into a shared product development workflow.
The new features aim to help teams move from early concepts to working software without constantly switching between design tools, code editors and documentation platforms.
The move comes as software companies increasingly build AI agents into product development. Figma's approach centres on its collaborative canvas, allowing people and AI tools to work with the same visual context rather than operating in separate environments.
CEO and co-founder Dylan Field described the direction as an evolution of Figma's multiplayer canvas into a full-stack creative environment where teams and agents can work together from idea to launch.
- Code Layers Bring Working Prototypes to the Canvas
Figma is rolling out Code Layers in open beta, allowing teams to build, edit, explore and compare interactive, code-backed prototypes directly on the canvas.
Users can create these layers through Figma's AI agent, connect a GitHub repository or work with coding agents through MCP. Existing Figma designs can also be converted into interactive code layers.
The feature could reduce friction between design and engineering teams by letting them test interactions and compare alternatives earlier in development. However, production readiness will still depend on code quality, testing and integration with a team's existing engineering processes.
Figma also introduced Specs, a native Markdown editor for recording product requirements and implementation details. Teams can create standalone documents or work alongside designs, embedding canvas content, diagrams and code layers to give both colleagues and AI agents the context they need.
Together, the features aim to narrow the gap between what a product team designs and what developers ultimately build.
- Figma Adds Sarvam to Its MCP Ecosystem
Figma's expanded MCP integration will connect its design environment with AI tools, including Sarvam, OpenCode, Dia and Meta Muse. The update allows compatible agents to interact with Figma's canvas, create code layers, edit specifications and generate visual assets.
For Sarvam, the partnership creates an opportunity to bring its AI capabilities into product development workflows used by designers and engineering teams.
Sarvam co-founder and CEO Pratyush Kumar said the collaboration combines the company's work on AI for India with Figma's design context to help teams improve product quality, reduce rework and accelerate development.
Figma is also consolidating design libraries, code libraries and guidelines into a centralised design system, giving AI agents a consistent reference for components and established design patterns. The feature is currently in preview.
- Payload Integration Brings Real Data Into Prototypes
Another update connects Payload, the open-source content management system acquired by Figma in 2025, with Figma Make. Teams can use it to build applications with a backend, database, media storage, APIs and an admin panel, bringing content and data closer to the design and prototyping process.
This could help product teams test applications with realistic content earlier, while allowing marketers to update information without needing to modify the codebase.
The broader strategy is clear: Figma wants to become more than a place to design interfaces. By bringing code, documentation, data and AI agents onto a shared platform, it is seeking a larger role in how digital products are built and maintained.
For developers and startups, the benefits will depend on how well these tools fit existing workflows, maintain code quality and preserve human control over production releases.
