oc mcp
Allows AI assistants and CLIs to read and modify the currently open Figma file through the Figma plugin API, including inspecting document info, components, and styles; getting selections; creating frames; adding elements; and posting design changes back.
This server does not currently have a deployable release on Glama. Follow the repository's setup instructions to install and configure it for your MCP client, including any required credentials and local dependencies.
figcontext — Figma ↔ AI bridge (no usage limits)
A simple file that lets your CLI know what's going on in your Figma file using the Figma plugin API. The plugin and bridge themselves are named oc mcp.
Let opencode, Claude Code, Cursor, or any AI read your open Figma file and change it. No Figma Dev Mode MCP, no 6/month limit.
The plugin auto-connects — no codes, no channels, nothing to paste.
Setup — 3 steps (2 minutes, first time only)
1. Install Node.js — get LTS from https://nodejs.org (skip if you already have it).
2. Start the bridge — either double-click start-bridge.bat (keeps a window open, says [bridge] listening ws://localhost:3055), or run bridge\ensure-bridge.vbs once for a hidden background start. After editing server.js, restart the bridge — Node loads the file once at startup.
3. Add the plugin to Figma — in Figma Desktop open any file:
Plugins → Development → Import plugin from manifest…→ pickmanifest.jsonin THIS folder (re-import only ifmanifest.jsonitself changes)Then
Plugins → Development → oc mcp→ it connects by itself and stays connected while its window is open
Done. Restart opencode/Claude/Cursor once so it loads the bridge (node setup-opencode.js already merged the MCP config).
Port
3055is automatic — only change it if that port is busy. If two bridges end up on one port, the second becomes a proxy instead of crashing — keep just one open to avoid confusion.
Related MCP server: OneWright
Everyday use — 3 steps (30 seconds)
1. Make sure the bridge is up (http://localhost:3055/health says ok).
2. In Figma run oc mcp — it auto-connects (keep its window open; it reconnects by itself if the connection drops).
3. Paste the link (http://localhost:3055/file) to ANY AI and tell it what to do:
this is my Figma file: <link> — code the hero section in React<link> — add a red button under the headline<link> — what components and styles are in this file?
The AI reads the live file from the link and posts changes back. Re-fetch the link for fresh state — never work from memory.
Power (no link, MCP tools): in opencode just say connect (or get_status, then get_document_info) — then get my selection and code it, create a frame, etc. 157 tools, same names over HTTP (POST /api/<command>) and MCP.
Share a public link (browser AI like ChatGPT/Claude.ai)
Local links only work on your PC. To share:
Make sure the bridge + Figma plugin are running.
Get
cloudflared.exeonce fromhttps://developers.cloudflare.com/cloudflare-one/connections/connect-networks/downloads/→ drop it in this folder.Double-click
share-link.bat→ keep open → it printshttps://something.trycloudflare.com— your share link isTHAT-URL/file. Paste to Claude/ChatGPT in browser. Closeshare-link.batto kill the link instantly.
Browser AIs are READ-only via the public link. Writing still needs the local bridge + CLI.
If it doesn't work
You see | Fix |
| Bridge not running — start it (bat or |
Link says | Plugin window closed — re-run plugin (it auto-connects) |
| Plugin window was closed mid-task — re-open plugin, retry once |
| Bridge is running old |
Port 3055 busy | Close extra bridge windows — only ONE at a time |
Copy button does nothing | Drag-select the text, |
Still stuck? Send: (1) last 10 lines of bridge output, (2) plugin log last 10 lines, (3) AI error text.
Manual / advanced (you don't need this)
npm --prefix bridge install
npm install
node node_modules/typescript/bin/tsc -p tsconfig.json
node setup-opencode.js # global config
node setup-opencode.js --project .\my-projectThe bat does all of this automatically. Only use manual steps for CI or custom paths. (npx is blocked by PowerShell execution policy — call node/npm.cmd directly.)
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
This server cannot be deployed
Maintenance
Related MCP Connectors
The Figma MCP server brings Figma design context directly into your AI workflow.
Talk to your public-facing AI from any MCP client — Claude, ChatGPT, Cursor, Cline, Windsurf.
Official Fibery MCP server: query and update your workspace from any AI client via OAuth.
Zero-setup MCP gateway securely connecting AI to your tools with authentication and workflows
Related MCP Servers
- AlicenseNot gradedqualityBmaintenanceLocal MCP server exposing Figma REST API tools to AI agents, enabling file reads, comments, variables, and other resource operations. Works with personal access tokens and integrates with Claude, Cursor, Codex, and more.MIT
- AlicenseNot gradedqualityAmaintenanceEnables AI agents to read and write Figma designs through a local WebSocket relay, turning Figma selections into framework-aware code and building/editing designs directly on the canvas. Provides 112 MCP tools for bidirectional design-code workflows with support for any MCP client.MIT
- AlicenseNot gradedqualityCmaintenanceEnables MCP-compatible AI clients to read and modify the user's currently open Figma file by executing JavaScript in Figma's sandbox, all through a local bridge with status monitoring, node jumping, and automatic rollback on errors.2MIT
- AlicenseAqualityCmaintenanceThe Figma desktop app as an MCP server. Gives AI agents the full Figma Plugin API on your open file: read, write, tokens, components, exports, undo. Runs locally with no cloud API, no token and no quota.1075 npm2MIT