Figma Make

Figma Make is an AI-powered application builder inside Figma that converts a plain-language description — or an existing Figma design frame — into a working, editable web application. It sits alongside Figma's canvas tools, so teams can move from a visual design to an interactive prototype without leaving the platform.^[1]

Background

Figma introduced Make in 2025 as generative-AI features spread from code assistants to design tooling. Rather than producing only static mockups, Make generates functional front-end code that can be previewed, edited and published, blurring the line between design files and shipped software. It belongs to the wave of AI app builders — along with Lovable, v0 and Bolt.new — that let non-developers produce working applications from prompts.^[2]

Key Features

• Natural-language prompting. Users describe what the app should do and how it should look; Make generates layouts, flows and functional components.^[2]
• Start from a frame. An existing Figma frame can be used as the visual starting point, keeping generated output aligned with the design system.
• Dual editing. Output can be refined either in a visual builder or in an embedded code editor for finer control over logic and layout.^[1]
• Data and auth. Make can connect to Supabase for user authentication, stored data and private APIs, and can scaffold a real backend with database, APIs and an admin panel.^[3]
• Credits and plans. AI usage is metered in credits, with allowances that vary by plan — Figma lists 4,250 credits per month on the Enterprise plan — and publishing apps requires a paid plan.^[3]
• Agent integration. Figma exposes MCP connectors so coding agents can read design context and make edits across Design, FigJam, Slides and Make files.^[3]

Applications

Make is used for dashboards, internal tools, customer portals, prototypes and small web apps — the "proof of concept by Friday" class of work that previously required a developer. It is most effective as a prototyping and internal-tool layer; production systems with complex requirements still typically need engineers to review the generated code, handle security and integrate with real backends.^[2]

>See Also

• Vibe Coding
• Lovable
• v0
• Cursor Code Editor
• Agentic AI
• AI and Copyright