모델 컨텍스트 프로토콜(MCP)을 사용하여 커서에 대한 Draw Things API 통합.
노드.js >= 14.0.0
http://127.0.0.1:7888 에서 실행되는 Draw Things API
Related MCP server: Jira Context MCP Server
지엑스피1
Cursor에서 이 도구를 설정하려면 cursor-setup.md 에서 자세한 가이드를 참조하세요.
빠른 설정:
~/.cursor/claude_desktop_config.json 생성하거나 편집합니다.
{
"mcpServers": {
"draw-things": {
"command": "draw-things-mcp-cursor",
"args": []
}
}
}커서 재시작
커서에서 사용: generateImage({"prompt": "a cute cat"})
echo '{"prompt": "your prompt here"}' | npx draw-things-mcp-cursorprompt : 이미지 생성을 위한 텍스트 프롬프트(필수)
negative_prompt : 이미지 생성을 위한 부정 프롬프트
width : 이미지 너비(기본값: 360)
height : 이미지 높이(기본값: 360)
steps : 생성을 위한 단계 수 (기본값: 8)
model : 생성에 사용할 모델(기본값: "flux_1_schnell_q5p.ckpt")
sampler : 샘플링 방법(기본값: "DPM++ 2M AYS")
예:
echo '{
"prompt": "a happy smiling dog, professional photography",
"negative_prompt": "ugly, deformed, blurry",
"width": 360,
"height": 360,
"steps": 4
}' | npx draw-things-mcp-cursorCursor에서 MCP 도구로 사용되는 경우 해당 도구는 다음 매개변수와 함께 generateImage 로 등록됩니다.
{
prompt: string; // Required - The prompt to generate the image from
negative_prompt?: string; // Optional - The negative prompt
width?: number; // Optional - Image width (default: 360)
height?: number; // Optional - Image height (default: 360)
model?: string; // Optional - Model name
steps?: number; // Optional - Number of steps (default: 8)
}생성된 이미지는 <sanitized_prompt>_<timestamp>.png 파일 이름 형식으로 images 디렉토리에 저장됩니다.
성공:
{
"type": "success",
"content": [{
"type": "image",
"data": "base64 encoded image data",
"mimeType": "image/png"
}],
"metadata": {
"parameters": { ... }
}
}오류:
{
"type": "error",
"error": "error message",
"code": 500
}문제가 발생하는 경우:
Draw Things API가 http://127.0.0.1:7888 에서 실행 중인지 확인하세요.
Cursor와 함께 사용하는 경우 ~/.cursor/logs 에 있는 로그 파일을 확인하세요.
src/index.js에 실행 권한이 있는지 확인하세요: chmod +x src/index.js
MIT
Generate an image based on a prompt
| Name | Required | Description | Default |
|---|---|---|---|
| prompt | No | ||
| negative_prompt | No | ||
| width | No | ||
| height | No | ||
| steps | No | ||
| seed | No | ||
| guidance_scale | No | ||
| random_string | No |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
With no annotations provided, the description carries the full burden of behavioral disclosure. It only states the action ('Generate an image') without mentioning any behavioral traits like rate limits, authentication needs, output format, or potential side effects (e.g., resource usage). For a tool with 8 parameters and no annotation coverage, this is inadequate, as it leaves critical operational details unspecified.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
The description is extremely concise with a single sentence, 'Generate an image based on a prompt', which is front-loaded and wastes no words. It efficiently communicates the core function without unnecessary elaboration, making it easy to parse quickly. This is optimal for brevity, though it may sacrifice detail.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
Given the complexity (8 parameters, no annotations, no output schema), the description is incomplete. It doesn't explain the return values, error conditions, or how parameters interact, leaving significant gaps for the agent to operate effectively. While conciseness is high, it lacks the necessary context to handle the tool's full scope, making it insufficient for robust use.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
The description adds no meaning beyond the input schema, which has 8 parameters and 0% schema description coverage. Parameters like 'negative_prompt', 'steps', and 'guidance_scale' are undocumented in both the schema and description, leaving their purposes unclear. This fails to compensate for the low coverage, resulting in poor parameter understanding for the agent.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
The description 'Generate an image based on a prompt' clearly states the verb ('Generate') and resource ('image'), making the purpose understandable. However, it's somewhat vague about the specific method or technology (e.g., AI model type), and with no siblings, it doesn't need differentiation but lacks depth. This meets the minimum viable standard without being tautological or misleading.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
The description provides no guidance on when to use this tool, such as scenarios for image generation, alternatives, or prerequisites. It merely states the basic function without context, leaving the agent to infer usage. This is a clear gap, as it doesn't help distinguish from potential implicit alternatives or set expectations.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
generateImageScored across 1 tool
With only one tool, there is no possibility of confusion or overlap between tools. The single tool 'generateImage' has a distinct and clear purpose, so disambiguation is perfect.
Since there is only one tool, naming consistency is inherently perfect. The tool name 'generateImage' follows a clear verb_noun pattern, and there are no other tools to compare it against for inconsistency.
A single tool for an image generation server feels thin and limited in scope. While it covers the core function, it lacks supporting operations like editing, listing, or managing images, making it borderline too few for a typical domain.
The server's purpose appears to be image generation, but with only a generate tool, there are significant gaps. There is no way to retrieve, update, delete, or manage generated images, which will likely cause agent failures in workflows requiring more than one-off generation.
Generate AI images, video, speech, music and presentations from Claude, ChatGPT and Cursor.
Image and video AI tools and your own pipelines, run from any AI assistant.
Generate images with your own ChatGPT subscription (Plus, Pro or Team), without spending API credits
Generate AI images, video, voiceovers and music from Claude, ChatGPT or Cursor through 50+ models (Veo 3.1, Kling 3, Seedance, Nano Banana, GPT Image, ElevenLabs). Also image editing, upscaling, background removal, face swap, transcription, voice cloning and UGC-style video ads. Sign in with OAuth — no API key to paste. Tools are annotated (read-only vs. credit-spending); failed generations are refunded.
