Skip to main content
Glama

What Is This?

LemGen AI Design MCP turns an AI coding agent into a practical visual creative assistant. Instead of asking the agent to invent prompts from nothing, it gives the agent a searchable prompt library, prompt enhancement tools, model metadata, local preferences, and optional image/video generation through LemGen.

Use it when you want your agent to:

  • find visual references before writing a prompt;

  • inspect full prompt examples with images, videos, model names, tags, and source links;

  • translate, polish, or improve prompts across multiple languages;

  • expand rough ideas into production-ready image or video prompts;

  • generate images or videos only after the user approves the final prompt.

Free tools work without an API key. Generation tools require LEMGEN_API_TOKEN from LemGen.

Related MCP server: vision-mcp

Preview

Quick Start

Install directly from GitHub today:

{
  "mcpServers": {
    "lemgen": {
      "command": "npx",
      "args": ["-y", "github:aithink001/lemgen-ai-design-mcp"]
    }
  }
}

Enable generation by adding your LemGen token:

{
  "mcpServers": {
    "lemgen": {
      "command": "npx",
      "args": ["-y", "github:aithink001/lemgen-ai-design-mcp"],
      "env": {
        "LEMGEN_API_TOKEN": "sk_your_key_here"
      }
    }
  }
}

After the npm package is published, the same server can be installed with:

npx -y lemgen@latest

Features

Capability

Tool

Free

What the agent gets

Prompt search

search_gallery

Yes

Ranked image/video prompt results with preview media and LemGen links.

Full inspiration

get_inspiration

Yes

Complete prompt, model, tags, author, stats, images/videos, and source URL.

Prompt expansion

enhance_prompt

Yes

A structured production prompt from a rough visual idea.

Model guidance

list_models

Yes

Supported models, media type, ratios, resolutions, reference limits, and strengths.

Prompt translation

prompt_tools

Yes

Translation, polish, and improvement workflows for image/video prompts.

Local memory

manage_preferences

Yes

Style, model, ratio, resolution, and favorite prompt IDs stored locally.

Image generation

generate_image

Token

LemGen image generation job with status and result URLs.

Video generation

generate_video

Token

LemGen video generation job; agents are instructed to ask before calling it.

Language Support

LemGen MCP is built for multilingual creative teams. Agent replies should follow the user's language, while model names and API parameters stay in English.

prompt_tools.targetLanguage supports:

Code

Language

Code

Language

Code

Language

en

English

zh

Chinese

ja

Japanese

ko

Korean

es

Spanish

fr

French

de

German

pt

Portuguese

it

Italian

nl

Dutch

ru

Russian

ar

Arabic

hi

Hindi

id

Indonesian

vi

Vietnamese

th

Thai

tr

Turkish

pl

Polish

Examples:

{ "action": "translate", "targetLanguage": "ja", "mediaType": "image" }
{ "action": "polish", "mediaType": "video" }

More setup notes are in docs/clients.md and docs/languages.md.

Client Setup

Claude Code

Install from this GitHub repository:

/plugin marketplace add aithink001/lemgen-ai-design-mcp
/plugin install lemgen@lemgen-marketplace

Restart Claude Code after installation.

Cursor

Add to Cursor MCP config:

{
  "mcpServers": {
    "lemgen": {
      "command": "npx",
      "args": ["-y", "github:aithink001/lemgen-ai-design-mcp"],
      "env": {
        "LEMGEN_API_TOKEN": "sk_your_key_here"
      }
    }
  }
}

Codex

codex mcp add lemgen -- npx -y github:aithink001/lemgen-ai-design-mcp

For generation, add LEMGEN_API_TOKEN to the MCP server environment in your Codex config.

Windsurf / Roo Code / Cline

Use the same MCP server block:

{
  "command": "npx",
  "args": ["-y", "github:aithink001/lemgen-ai-design-mcp"],
  "env": {
    "LEMGEN_API_TOKEN": "sk_your_key_here"
  }
}

OpenClaw

Install from ClawHub:

clawhub install lemgen-ai-design

Skill page:

https://clawhub.ai/aithink001/skills/lemgen-ai-design

The source skill is included in this repository:

openclaw/SKILL.md

It defines the recommended workflow for inspiration search, prompt enhancement, generation confirmation, and video safety.

Prompt Library

This repository ships with data/trending-prompts.json, exported from LemGen.

Type

Count

Image prompts

2,593

Video prompts

1,940

Total

4,533

Model

Count

GPT Image

1,935

Seedance 2.0

1,940

Nano Banana Pro

489

Midjourney

104

Other image models

65

Schema:

type GalleryPrompt = {
  id: string;
  rank: number;
  title: string;
  prompt: string;
  author: string;
  model: string;
  categories: string[];
  image: string;
  images: string[];
  video_url?: string;
  media_type: "image" | "video";
  likes: number;
  views: number;
  source_url: string;
};

Data-only repository:

https://github.com/aithink001/lemgen-trending-prompts

Development

pnpm install
pnpm validate
pnpm inspect

pnpm validate runs TypeScript, build, brand scan, and an MCP smoke test.

Environment Variables

Variable

Default

Description

LEMGEN_API_BASE_URL

https://lemgen.org

LemGen API host.

LEMGEN_API_TOKEN

none

Required for image/video generation.

LEMGEN_OUTPUT_DIR

~/Pictures/lemgen

Reserved for future CLI image saving.

LEMGEN_VIDEO_OUTPUT_DIR

~/Movies/lemgen

Reserved for future CLI video saving.

LEMGEN_PREFERENCES_PATH

~/.lemgen/preferences.json

Local preferences file.

Principles

  • Search and inspiration are free; generation is opt-in.

  • Video generation requires explicit confirmation.

  • Responses return real IDs and URLs, not guessed visual descriptions.

  • Public data is source-linked to LemGen.

  • Documentation and package metadata stay clean of duplicate-brand signals.

Roadmap

  • Publish the lemgen npm package.

  • Add a hosted remote MCP endpoint.

  • Add semantic search over the prompt library.

  • Add focused prompt packs for product photography, logos, portraits, posters, and video ads.

  • Add one-shot CLI generation commands for shell and CI workflows.

License

MIT

Available Tools

8 tools
enhance_promptA
Read-only

Turn a short idea into a professional image/video prompt. Free, local, no API key required.

Inputs: idea (required), style, mediaType, aspectRatio, +1 more.

generate_imageA

Generate an AI image with LemGen. Requires LEMGEN_API_TOKEN. Use after the user approves a final prompt.

Inputs: prompt (required), wait, model, quality, +5 more.

generate_videoB

Generate an AI video with LemGen. Requires LEMGEN_API_TOKEN. Videos are slow and should only be called after explicit user confirmation.

Inputs: prompt (required), tier, wait, model, +7 more.

get_inspirationA
Read-only

Get the full prompt, preview media, metadata, and LemGen source link for one gallery entry. Free, no API key required.

Inputs: id (required).

list_modelsA
Read-only

List LemGen image and video models with strengths, ratios, resolutions, and reference limits. Free, no API key required.

Inputs: mediaType.

manage_preferencesA

Read or update local LemGen preferences for style, model, aspect ratio, resolution, and favorite prompt ids. Free, local, no API key required.

Inputs: action (required), model, style, promptId, +3 more.

prompt_toolsC
Read-only

Translate, polish, or improve an AI image/video prompt. Free locally; if LEMGEN_API_TOKEN is configured, the LemGen prompt tool may be used by callers separately.

Inputs: text (required), action (required), mediaType, targetLanguage.

Tool Schema Changelog

Recent tool additions, removals, and schema changes observed during successful MCP inspections.

Loading tool schema changelog…

    TDQS

    A3.7/5.0

    Scored across 8 tools

    Disambiguation4/5

    Most tools have distinct purposes, but 'enhance_prompt' and 'prompt_tools' both deal with prompt improvement, which could cause confusion despite different specific actions.

    Naming Consistency5/5

    All tool names follow a consistent verb_noun pattern with lowercase and underscores, e.g., 'generate_image', 'list_models', 'search_gallery'.

    Tool Count5/5

    8 tools is well-scoped for an AI design MCP, covering prompt enhancement, generation, inspiration, models, preferences, and gallery search without being excessive.

    Completeness4/5

    The tool set covers the core workflow of prompt creation, generation, and exploration, but lacks tools for managing generated assets (e.g., deletion or history), which is a minor gap.

    Maintenance

    ActivityInactive
    ResponsivenessNo recent community issues

    Related MCP Connectors

    Related MCP Servers