claude-talk-to-figma-mcp and figma-mcp-server
These are competitors offering similar MCP server implementations for Figma integration, though A has significantly broader adoption and support for multiple AI clients while B appears to be a more minimal local-only implementation.
About claude-talk-to-figma-mcp
arinspunk/claude-talk-to-figma-mcp
A Model Context Protocol (MCP) that allows Claude Desktop and other AI tools (Claude Code, Cursor, Antigravity, etc.) to read, analyze, and modify Figma designs
Implements a WebSocket-based plugin architecture that bridges Figma Desktop with MCP servers, enabling real-time bidirectional communication without requiring a Dev Mode license. Supports design analysis (contrast audits, component scanning), bulk modifications (colors, typography, auto-layout), and code generation (React/Vue/SwiftUI components with design tokens). Integrates across multiple AI agents including Claude Desktop, Cursor, Windsurf, and VS Code extensions via configurable MCP server setup.
About figma-mcp-server
planetabhi/figma-mcp-server
A comprehensive local MCP server for Figma
Exposes Figma design data through multiple MCP tool endpoints (like `get_file_nodes`) that parse design node links and return structured metadata. Built on Bun runtime with stdio transport, it integrates directly into Claude Desktop, Cursor, and Gemini CLI via configuration files, requiring only a Figma API key for authentication. Supports both default stdio mode and optional SSE/web server modes for flexible deployment scenarios.
Related comparisons
Scores updated daily from GitHub, PyPI, and npm data. How scores work