Overview
The Framer MCP plugin (officially titled "MCP: AI Plugin" by Tommy D. Rossi) is a community-developed but widely adopted tool available on the Framer Marketplace. It implements an MCP server endpoint that allows any Model Context Protocol-compatible AI client to interact directly with open Framer projects. Launched around mid-2025, it has become a go-to solution for designers and developers wanting to bring agentic AI workflows into Framer's visual canvas.
By establishing a secure WebSocket tunnel routed through a Cloudflare Worker, the plugin exposes Framer's Plugin API to AI agents, enabling real-time canvas operations without manual copy-paste or custom scripting.
Key Features
- Secure Tunnel Connection: Uses your Framer user ID for authentication and routes requests via Cloudflare Worker to your open plugin instance.
- Canvas Manipulation: Create frames, add text/layers/images, modify styles, reorder/delete elements, zoom to selections — all via natural language.
- Content & Automation: Generate AI-written copy, populate CMS fields, apply design updates automatically.
- React Code Export: Convert selected Framer components or entire pages to clean React/TSX code (often paired with Unframer CLI).
- Multi-Client Support: Works seamlessly with Claude Desktop/Code, Cursor, GitHub Copilot (MCP mode), and other MCP clients.
- No Server Management: Runs in-browser via the plugin; no need to host your own MCP server.
How it works technically: The plugin acts as an MCP server proxy. AI clients connect to a provided URL (e.g., via claude mcp add "framer-mcp" <url>), send JSON-RPC calls over the tunnel, and the plugin executes actions using Framer's runtime API.
Installation & Quick Start
- Open your Framer project.
- Go to Plugins → Marketplace → Search for "MCP" or visit https://www.framer.com/marketplace/plugins/mcp.
- Install and open the plugin in your project (must remain open for the tunnel to stay active).
- Copy the MCP server URL displayed in the plugin widget.
- Add it to your AI client:
- Claude:
claude mcp add "framer-mcp" <your-url> - Cursor/VS Code: Add to
~/.cursor/mcp.jsonor equivalent settings.
- Claude:
- Start prompting: "Create a hero section with a heading 'Welcome' and a blue button" or "Export this frame as React component".
Use Cases
- Rapid Prototyping: Ask AI to generate layouts, iterate on designs conversationally.
- Content Population: Auto-fill marketing copy, product descriptions, or CMS entries from prompts.
- Design-to-Code: Export polished Framer components to React for Next.js/Vercel deployment.
- Automation Workflows: Update multiple pages with new branding, responsive adjustments, or seasonal content.
- Collaborative AI Assistance: Designers use Claude to explore ideas while keeping full visual control in Framer.
Community feedback from 2026 shows 3–6× faster iteration cycles for Framer users leveraging this plugin compared to traditional manual edits or screenshot-based AI prompting.
Advanced Tips & Common Pitfalls
- Keep the plugin window open in Framer — closing it terminates the tunnel.
- Use specific, layered prompts for complex changes (e.g., "Apply dark mode to all frames using #111 background and #eee text").
- Combine with Unframer (https://github.com/remorses/unframer) for CLI-based exports and bulk operations.
- Pitfall: Tunnel security relies on Framer auth — avoid sharing URLs publicly.
- Enterprise-tip: For always-on access, explore Framer's Server API alongside custom MCP servers for headless updates.
Related Tools & Ecosystem
- Official Framer MCP Dictionary entries: https://www.framer.com/dictionary/model-context-protocol
- Unframer CLI for enhanced code export: https://github.com/remorses/unframer
- Design Bridge MCP (alternative plugin with more canvas utilities): https://www.framer.com/marketplace/plugins/design-bridge-mcp
- Framer Server API for production/headless publishing: https://www.framer.com/developers/server-api-introduction
The Framer MCP plugin exemplifies how visual design tools are evolving into agent-native platforms in 2026. Install it today to experience AI truly collaborating inside your Framer canvas.