I’ve heard a few people mention that the Figma Console MCP setup feels cumbersome or intimidating at first glance. That reaction makes sense. Any workflow that touches local setup, terminals, config files, and Figma permissions can look heavier than it actually is.
So I recorded a full walkthrough and timed it.
The video above shows the full local installation, start to finish, with no steps skipped. From clone to a live MCP connection inside Figma takes just a few minutes, and you can see exactly what’s involved before deciding whether it’s something you want to run locally.
Two Setup Options, Two Different Tradeoffs
There are two primary ways to run the Figma Console MCP.
The remote setup is the lightest-weight option. You provide a URL and a name, and you immediately get access to analysis and reporting features. This is useful if you want to explore what the MCP can “see” without touching your local environment.
The local setup is what unlocks everything. Read/write access. Deeper inspection. Full interaction with tokens, components, and metadata. This is the path most people want once they move past initial curiosity.
This walkthrough covers the full-featured local setup, which is what unlocks write access, deeper inspection, and full interaction with tokens, components, and metadata. There’s also a lighter remote option if you just want to explore analytics and reporting, but the local setup is where the real power lives.
A couple of things worth clarifying up front:
This is a one-time setup, not something you repeat per Figma file or per codebase
Once running, the MCP can work across multiple Figma projects and multiple repositories
The goal isn’t to memorize setup steps, but to get to a place where the tooling disappears and you just use it
If you watch the video and want more detail, the documentation goes deeper on every piece of the setup, including alternate install paths and configuration options:
And if you run into friction anyway, reach out. I’m happy to help get people unblocked. The setup shouldn’t be the thing that prevents you from seeing whether this fits your workflow.
Once it’s running, the interesting work starts.
