import { Server } from "@modelcontextprotocol/sdk/server/index.js";
import { StdioServerTransport } from "@modelcontextprotocol/sdk/server/stdio.js";
import {
CallToolRequestSchema,
ListToolsRequestSchema,
} from "@modelcontextprotocol/sdk/types.js";
import sharp from "sharp";
import fetch from "node-fetch";
import fs from "fs/promises";
import os from "os";
import path from "path";
const DOG_API_URL = "https://dog.ceo/api/breeds/image/random";
async function fetchRandomDogImage(): Promise<Buffer> {
const response = await fetch(DOG_API_URL);
const data = await response.json() as { message: string; status: string };
if (data.status !== "success") {
throw new Error("Failed to fetch dog image");
}
const imageResponse = await fetch(data.message);
const buffer = await imageResponse.buffer();
return buffer;
}
async function addLGTMOverlay(imageBuffer: Buffer): Promise<Buffer> {
// Get image metadata
const metadata = await sharp(imageBuffer).metadata();
const imageWidth = metadata.width || 800;
const imageHeight = metadata.height || 600;
// Calculate text size based on image dimensions
const fontSize = Math.floor(Math.min(imageWidth, imageHeight) / 6);
// Create SVG text with white text and black stroke
const svg = `
<svg width="${imageWidth}" height="${imageHeight}">
<text x="50%" y="50%"
dominant-baseline="middle"
text-anchor="middle"
font-family="Arial, sans-serif"
font-size="${fontSize}"
font-weight="normal"
fill="white"
stroke="black"
stroke-width="${fontSize / 25}"
style="paint-order: stroke fill;">
LGTMだワン！
</text>
</svg>
`;
// Composite the text overlay onto the image
const outputBuffer = await sharp(imageBuffer)
.composite([
{
input: Buffer.from(svg),
gravity: 'center'
}
])
.png()
.toBuffer();
return outputBuffer;
}
const server = new Server(
{
name: "lgtm-dog-mcp",
version: "1.0.0",
},
{
capabilities: {
tools: {},
},
}
);
server.setRequestHandler(ListToolsRequestSchema, async () => {
return {
tools: [
{
name: "generate_lgtm_dog",
description: "Generate a dog image with LGTM overlay",
inputSchema: {
type: "object",
properties: {
outputPath: {
type: "string",
description: "Path where the generated image should be saved (optional)",
},
},
},
},
],
};
});
server.setRequestHandler(CallToolRequestSchema, async (request) => {
if (request.params.name === "generate_lgtm_dog") {
try {
const args = request.params.arguments as { outputPath?: string };
const dogImageBuffer = await fetchRandomDogImage();
const lgtmImageBuffer = await addLGTMOverlay(dogImageBuffer);
// Generate unique filename
const timestamp = Date.now();
const filename = args.outputPath || path.join(os.homedir(), 'Downloads', `lgtm-dog-${timestamp}.png`);
// Save to temporary location
await fs.writeFile(filename, lgtmImageBuffer);
return {
content: [
{
type: "text",
text: `LGTM dog image generated successfully!\n\nImage saved to: ${filename}\n\nYou can open this file to view the image.`,
},
],
};
} catch (error) {
return {
content: [
{
type: "text",
text: `Error generating LGTM dog image: ${error}`,
},
],
};
}
}
throw new Error(`Unknown tool: ${request.params.name}`);
});
async function main() {
const transport = new StdioServerTransport();
await server.connect(transport);
console.error("LGTM Dog MCP server running on stdio");
}
main().catch((error) => {
console.error("Server error:", error);
process.exit(1);
});
