How to Connect Figma to Gemini via MCP

💡 No install? Use cloud.anythingmcp.com directly. Sign in, click Connectors → Figma, paste your credentials, mint an MCP API key and you're done. No Docker, no git clone, no local server.

Figma + Gemini

Let your AI read your Figma designs: browse team folders and files, inspect pages, frames and layers, export images, look up components and styles, follow version history, read and post comments, and manage Dev Mode links. With a personal access token.

Prerequisites

See the full setup instructions baked into the connector (visible in the in-app store when you select the connector). The required environment variables for this connector are:

FIGMA_ACCESS_TOKEN

Step 1 — Get credentials

1. In Figma, open the account menu (top left in the file browser) → Settings → Security → Personal access tokens → Generate new token.
2. Name the token (for example "AnythingMCP") and choose an expiry: at most 90 days, as Figma no longer issues tokens that never expire. Set these scopes:
   • current_user:read, file_content:read, file_metadata:read, file_versions:read
   • file_comments:read, file_comments:write
   • library_content:read, library_assets:read, team_library_content:read
   • folders:read
   • file_dev_resources:read, file_dev_resources:write
3. Click Generate token and copy it right away: Figma shows it only once. It usually starts with figd_.
4. Paste the token into FIGMA_ACCESS_TOKEN, without the word Bearer, then install the connector.
5. Start with figma_get_me: it shows which Figma user the token belongs to. Then give the AI a file link: the file key is the part after /design/ in the URL.

Token expiry: tokens last at most 90 days. When yours expires, generate a new one with the same scopes and paste it into FIGMA_ACCESS_TOKEN again. The connector sees exactly the files and teams the token's user can open. Tokens made before Figma renamed projects to folders may only carry projects:read, while the folder tools need folders:read: if they answer 403 naming a scope, generate a new token.

Finding teams and folders: Figma cannot list your teams through its API. Give the AI the link of a team page from the file browser (the team id is the number after /team/); figma_list_team_folders, figma_list_subfolders and figma_list_folder_files then lead to the files.

Permissions and read-only setup: leave out file_comments:write and file_dev_resources:write, and switch off figma_post_comment, figma_delete_comment, figma_add_comment_reaction, figma_create_dev_resource and figma_update_dev_resource. Comments the AI posts appear under the token user's name, and everyone with access to the file can see them.

Switched off at install: figma_delete_dev_resource, which removes a Dev Mode link from a node for everyone. Turn it on only if the AI should delete dev resources.

Rate limits and the Starter plan: Figma limits requests per user and by the plan the file is on. Reading file content, nodes and image renders (figma_get_file, figma_get_file_nodes, figma_render_images) has the tightest limit: 10 to 20 a minute on a paid plan with a Full or Dev seat, but only a handful a month for files on the free Starter plan or with a View or Collab seat. Metadata, comments, versions, components and styles allow far more. Ask the AI to look at single frames rather than whole files; very large answers are cut at 200 KB. Rendered image links stay valid for 30 days.

Dev Mode: dev resources (links to tickets, docs or code attached to a node) need a paid plan and a Full or Dev seat. Without Dev Mode access, Figma answers "File not found" to the dev resource tools, even for a file the other tools read fine. A node holds at most 10 dev resources, each URL once.

Common errors: 401 "Invalid token" means the token is wrong, expired or revoked: generate a new one. 403 means the token lacks a scope (the message names it) or the user has no access to that team or folder. 404 means there is no such file or node, or the user cannot open the file: copy the key from the file's URL. 400 names an invalid parameter, for example a node id with a hyphen instead of a colon (a link's node-id=12-345 is 12:345). 429 is a rate limit: wait and try again. The library tools refuse a branch key: use the main file's key.

The connector does not edit designs, and it does not handle variables, webhooks, FigJam and Slides content beyond the file tree, or library analytics.

Step 2 — Install the adapter

curl -fsSL https://raw.githubusercontent.com/HelpCode-ai/anythingmcp/main/docker-compose.quickstart.yml -o docker-compose.yml
printf 'JWT_SECRET=%s\nENCRYPTION_KEY=%s\n' "$(openssl rand -hex 32)" "$(openssl rand -hex 32)" > .env
docker compose up -d

Step 3 — Add the connector in Gemini

Your server URL: in AnythingMCP, open MCP Servers → the server this connector is on, and copy its URL (https://cloud.anythingmcp.com/mcp/…). Use it wherever this guide shows YOUR_SERVER_ID.

Gemini CLI reads MCP servers from ~/.gemini/settings.json (on Windows %USERPROFILE%\.gemini\settings.json). Add:

{
  "mcpServers": {
    "anythingmcp": {
      "httpUrl": "https://cloud.anythingmcp.com/mcp/YOUR_SERVER_ID",
      "headers": { "X-API-Key": "YOUR_MCP_API_KEY" }
    }
  }
}

1. Create an MCP API key in AnythingMCP under MCP Servers → your server → API keys, and paste it in place of YOUR_MCP_API_KEY.
2. Save the file and restart gemini.
3. Run /mcp inside Gemini CLI (or gemini mcp list in your shell): anythingmcp should show as connected.
4. Start prompting: Gemini CLI can now call the Figma tools.

Available tools

Example prompts

• "Show me the pages and main frames of the Checkout redesign file, and who edited it last."
• "Which comments on the Checkout redesign file are still unresolved, and who wrote them?"
• "Export the Pricing page frame of the Acme website file as a PNG and give me the link."
• "Reply to Lena's comment on the Checkout redesign file and tell her the button color has been updated."

FAQ

Where does Gemini CLI look for MCP servers? In ~/.gemini/settings.json for your user, or in .gemini/settings.json inside a project. gemini mcp list shows each server and whether it connects.

Next steps

• Connector landing page
• Connect to Claude
• Connect to ChatGPT
• Connect to Cursor
• Connect to Meta Muse