jau123/MeiGen-AI-Design-MCP

What it solves

MeiGen AI Design MCP transforms AI coding tools (like Cursor, Claude Code, and Windsurf) into professional design assistants. It bridges the gap between text-based coding environments and visual content creation by allowing developers to generate images and videos directly within their IDE or terminal without switching contexts.

How it works

It operates as a Model Context Protocol (MCP) server that provides a set of tools and slash commands to AI agents. It supports three backend modes for generation:

• MeiGen Cloud: A managed service with a wide lineup of image and video models.
• OpenAI-compatible APIs: Integration with third-party providers like Together AI or Fireworks AI.
• Local ComfyUI: Direct connection to a local GPU for offline, private generation using custom workflows.

The system uses specialized sub-agents (like a prompt-crafter and image-generator) to handle parallel batch generation, ensuring the main chat context remains clean while multiple visual variations are produced simultaneously.

Who it’s for

• Developers using AI coding assistants who need to create logos, product shots, or UI assets on the fly.
• Designers who want to automate visual iterations using LLM-driven prompt engineering.
• Power users who prefer running local generative AI via ComfyUI but want an LLM interface to control it.

Highlights

• Multi-Backend Support: Seamlessly switch between cloud APIs and local ComfyUI instances.
• Curated Prompt Library: Access to over 1,400 trending prompt templates for high-quality inspiration.
• Parallel Orchestration: Generates multiple creative directions in parallel using sub-agents.
• Broad Compatibility: Works across major MCP-compatible hosts including Claude Code, Cursor, Windsurf, and Roo Code.
• Standalone CLI: Includes a gen command for use in shell scripts and CI pipelines without needing an MCP host.