Skip to main content
Glama
emanirfan-creator

Pallas MCP Server

Pallas MCP Server

A Node.js MCP (Model Context Protocol) server that connects an AI agent to both Figma and a Panda CSS codebase simultaneously to handle two-way design system sync, component creation, and variable binding.

Features

  • Figma Variable Management: Read/write primitives and aliases.

  • Auto Layout & Component Creation: Construct full variant sets systematically.

  • Panda CSS Syncing: Push recipes/tokens to Figma & pull updates to Code.

  • Token Structure Enforcer: Validate semantic-to-primitive maps across states.

Related MCP server: Agent to Figma MCP

Installation

  1. Install dependencies for the MCP Server:

cd pallas-mcp
npm install
npm run build
  1. Open Figma, load your design file.

  2. Import the Companion Plugin into Figma:

    • Go to Plugins -> Development -> Import plugin from manifest...

    • Choose pallas-mcp/figma-plugin/manifest.json.

Usage

  1. Start the MCP Server: Typically, this process is spawned by the MCP host (like the AI agent). Provide its path in the MCP configuration:

{
  "mcpServers": {
    "pallas-mcp": {
      "command": "node",
      "args": ["/absolute/path/to/pallas-mcp/dist/index.js"]
    }
  }
}
  1. Run the Companion Plugin: In Figma, launch the Pallas MCP Companion plugin. It opens a WebSocket listener on ws://localhost:3055. The plugin must be active for the MCP Server to modify the Figma document.

  2. Initialize Design System Map: To establish a baseline mapping and track syncs, ensure /design-map.json is created:

{
  "tokens": {},
  "components": {}
}

Workflows

Add a Component End-To-End

The AI Agent operates these MCP tools to establish a structured component:

  1. figma_ensure_token_structure (Validate Recipe dependencies)

  2. panda_parse_recipe_to_figma_schema (Map structure)

  3. figma_create_component (Base Setup)

  4. figma_add_auto_layout (Layers Configuration)

  5. figma_create_component_set (Generate variants)

  6. figma_add_component_property (Wire props)

  7. figma_resolve_recipe_token_map

  8. figma_bind_all_variant_tokens

  9. figma_verify_component_bindings

  10. map_write (Commit map)

Change a Token

Ask the Agent to:

  1. panda_write_token / figma_write_variable

  2. sync_push & sync_pull for parity testing.

Troubleshooting

  • Connection Error / Timeouts: Ensure the Figma Plugin panel is open. The plugin communicates via WebSockets locally.

  • Unbound Properties (figma_verify_component_bindings fails): Do not use raw colors. Update the target layer styles using figma_bind_variable_to_property.

Related MCP Connectors

Related MCP Servers

  • A
    license
    Not graded
    quality
    D
    maintenance
    Connects Figma designs to AI agents, enabling extraction of production-ready code, assets, and design tokens through natural language descriptions. Supports React, Vue, CSS, and Tailwind with real-time design system analysis.
    55,619 npm
    40
    MIT
  • F
    license
    A
    quality
    D
    maintenance
    Enables AI agents to interact with Figma to create, read, and manage designs using the Figma REST API and a dedicated plugin. It supports advanced features like UI generation from text, webpage reconstruction in Figma, and design token synchronization with codebases.
    20
    -
  • A
    license
    Not graded
    quality
    D
    maintenance
    Enables AI assistants like Claude and Cursor to interact directly with Figma to create, modify, and analyze design elements in real-time. It provides a comprehensive suite of tools for document inspection, styling, component management, and automated layout via a bidirectional WebSocket connection.
    165 npm
    MIT