Search
Connect your own agents to Framer MCP
Design, build, and update your site with AI tools like ChatGPT, Claude, and Cursor, without giving up the flexibility, reliability, and control of Framer.
Using an app that’s not listed? Follow the manual setup guide and use this server URL.
https://mcp.framer.com/mcp
https://mcp.framer.com/mcp
Learn how to use the Framer MCP
Work anywhere, stay in control
Work from your desktop, browser, or phone
Use Framer from any supported desktop, browser, or mobile client. There’s no local Framer bridge to install and no need to keep the editor open.
Review every change before it goes live
Changes stay in your project or branch until you or your agent publishes them. Review the work, compare it with main, and publish when you’re ready.
Bring Framer to your AI product
Connect your MCP client to Framer so users can create, edit, and manage Framer projects from your product.
Frequently asked questions
What is MCP?
MCP (Model Context Protocol) lets AI agents connect to apps like Framer. Once connected, your agent can work directly in your Framer projects to build pages, update content, and more.
Why use MCP instead of Framer’s built-in agent?
You can work from an agent you already use, on your desktop, in your browser, or on your phone where supported. Your Framer project doesn’t need to be open. Your AI usage follows that agent’s plan or credits.
What can my agent access?
After you approve the connection, your agent can work with the canvas, components, supported project settings and CMS in projects your Framer account has access to. It uses your account’s permissions.
Can my agent change my live site?
Yes, if you explicitly ask it to publish. By default, changes stay in your project and don’t affect the live site until you publish. Review the work in Framer or on a branch before publishing.
What if my agent makes a mistake?
You can ask your agent to work on a branch, so you can review its changes and fix anything that isn’t right before publishing.
Which AI tools work with Framer MCP?
Any AI tool that supports remote MCP servers and OAuth can connect to Framer, including Codex, Claude Code, Cursor, Pi, Windsurf, Warp, and Grok.
How is this different from the local CLI setup?
Framer MCP connects compatible AI tools through a remote MCP server. The existing @framer/agent setup uses a local CLI bridge and currently supports additional workflows, including localization. Both remain supported while Framer MCP reaches feature parity.