Nano Banana Pro MCP

#!/usr/bin/env node 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 { z } from "zod"; import { writeFile, mkdir } from "fs/promises"; import { dirname, resolve } from "path"; import { GeminiImageClient } from "./gemini.js"; async function saveImageToFile( base64Data: string, outputPath: string ): Promise<string> { const absolutePath = resolve(outputPath); const dir = dirname(absolutePath); // Ensure directory exists await mkdir(dir, { recursive: true }); // Decode base64 and write to file const buffer = Buffer.from(base64Data, "base64"); await writeFile(absolutePath, buffer); return absolutePath; } const imageInputSchema = z.object({ data: z.string().describe("Base64 encoded image data"), mimeType: z.string().describe("MIME type of the image (e.g., image/png, image/jpeg)"), }); const generateImageSchema = z.object({ prompt: z.string().describe("Description of the image to generate"), aspectRatio: z .enum(["1:1", "3:4", "4:3", "9:16", "16:9"]) .optional() .default("1:1") .describe("Aspect ratio of the generated image"), imageSize: z .enum(["1K", "2K", "4K"]) .optional() .default("1K") .describe("Resolution of the generated image"), model: z .string() .optional() .describe("Gemini model to use (default: gemini-3-pro-image-preview)"), images: z .array(imageInputSchema) .optional() .describe("Optional reference images to guide generation"), outputPath: z .string() .optional() .describe("Optional file path to save the generated image (e.g., /path/to/image.png)"), }); const editImageSchema = z.object({ prompt: z.string().describe("Instructions for how to edit the image(s)"), images: z .array(imageInputSchema) .min(1) .describe("One or more images to edit"), model: z .string() .optional() .describe("Gemini model to use (default: gemini-3-pro-image-preview)"), outputPath: z .string() .optional() .describe("Optional file path to save the edited image (e.g., /path/to/image.png)"), }); const describeImageSchema = z.object({ images: z .array(imageInputSchema) .min(1) .describe("One or more images to describe/analyze"), prompt: z .string() .optional() .describe("Optional custom prompt for analysis (default: general description)"), model: z .string() .optional() .describe("Gemini model to use (default: gemini-3-pro-image-preview)"), }); export function createServer(apiKey: string): Server { const client = new GeminiImageClient(apiKey); const server = new Server( { name: "nano-banana-pro-mcp", version: "1.0.0", }, { capabilities: { tools: {}, }, } ); server.setRequestHandler(ListToolsRequestSchema, async () => { return { tools: [ { name: "generate_image", description: "Generate an image using Google Gemini. Optionally provide reference images to guide the generation style or content. Returns a base64-encoded image.", inputSchema: { type: "object" as const, properties: { prompt: { type: "string", description: "Description of the image to generate", }, aspectRatio: { type: "string", enum: ["1:1", "3:4", "4:3", "9:16", "16:9"], description: "Aspect ratio of the generated image", default: "1:1", }, imageSize: { type: "string", enum: ["1K", "2K", "4K"], description: "Resolution of the generated image", default: "1K", }, model: { type: "string", description: "Gemini model (gemini-3-pro-image-preview, gemini-2.5-flash-preview-05-20, or gemini-2.0-flash-exp)", default: "gemini-3-pro-image-preview", }, images: { type: "array", description: "Optional reference images to guide generation", items: { type: "object", properties: { data: { type: "string", description: "Base64 encoded image data" }, mimeType: { type: "string", description: "MIME type (e.g., image/png)" }, }, required: ["data", "mimeType"], }, }, outputPath: { type: "string", description: "Optional file path to save the generated image (e.g., /path/to/image.png)", }, }, required: ["prompt"], }, }, { name: "edit_image", description: "Edit one or more images using Google Gemini. Provide images and instructions for how to modify them. Returns a base64-encoded image.", inputSchema: { type: "object" as const, properties: { prompt: { type: "string", description: "Instructions for how to edit the image(s)", }, images: { type: "array", description: "One or more images to edit", items: { type: "object", properties: { data: { type: "string", description: "Base64 encoded image data" }, mimeType: { type: "string", description: "MIME type (e.g., image/png)" }, }, required: ["data", "mimeType"], }, minItems: 1, }, model: { type: "string", description: "Gemini model (gemini-3-pro-image-preview, gemini-2.5-flash-preview-05-20, or gemini-2.0-flash-exp)", default: "gemini-3-pro-image-preview", }, outputPath: { type: "string", description: "Optional file path to save the edited image (e.g., /path/to/image.png)", }, }, required: ["prompt", "images"], }, }, { name: "describe_image", description: "Analyze and describe one or more images using Google Gemini. Returns a text description of the image contents.", inputSchema: { type: "object" as const, properties: { images: { type: "array", description: "One or more images to describe/analyze", items: { type: "object", properties: { data: { type: "string", description: "Base64 encoded image data" }, mimeType: { type: "string", description: "MIME type (e.g., image/png)" }, }, required: ["data", "mimeType"], }, minItems: 1, }, prompt: { type: "string", description: "Optional custom prompt for analysis (default: general description)", }, model: { type: "string", description: "Gemini model (gemini-3-pro-image-preview, gemini-2.5-flash-preview-05-20, or gemini-2.0-flash-exp)", default: "gemini-3-pro-image-preview", }, }, required: ["images"], }, }, ], }; }); server.setRequestHandler(CallToolRequestSchema, async (request) => { if (request.params.name === "generate_image") { try { const args = generateImageSchema.parse(request.params.arguments); const result = await client.generateImage(args); // Save to file if outputPath is provided let savedPath: string | undefined; if (args.outputPath) { savedPath = await saveImageToFile(result.base64Data, args.outputPath); } return { content: [ { type: "image" as const, data: result.base64Data, mimeType: result.mimeType, }, ...(savedPath ? [{ type: "text" as const, text: `Image saved to: ${savedPath}` }] : []), ...(result.description ? [{ type: "text" as const, text: result.description }] : []), ], }; } catch (error) { const message = error instanceof Error ? error.message : "Unknown error occurred"; return { isError: true, content: [ { type: "text" as const, text: `Failed to generate image: ${message}`, }, ], }; } } if (request.params.name === "edit_image") { try { const args = editImageSchema.parse(request.params.arguments); const result = await client.generateImage({ prompt: args.prompt, images: args.images, model: args.model, }); // Save to file if outputPath is provided let savedPath: string | undefined; if (args.outputPath) { savedPath = await saveImageToFile(result.base64Data, args.outputPath); } return { content: [ { type: "image" as const, data: result.base64Data, mimeType: result.mimeType, }, ...(savedPath ? [{ type: "text" as const, text: `Image saved to: ${savedPath}` }] : []), ...(result.description ? [{ type: "text" as const, text: result.description }] : []), ], }; } catch (error) { const message = error instanceof Error ? error.message : "Unknown error occurred"; return { isError: true, content: [ { type: "text" as const, text: `Failed to edit image: ${message}`, }, ], }; } } if (request.params.name === "describe_image") { try { const args = describeImageSchema.parse(request.params.arguments); const description = await client.describeImage({ images: args.images, prompt: args.prompt, model: args.model, }); return { content: [ { type: "text" as const, text: description, }, ], }; } catch (error) { const message = error instanceof Error ? error.message : "Unknown error occurred"; return { isError: true, content: [ { type: "text" as const, text: `Failed to describe image: ${message}`, }, ], }; } } return { isError: true, content: [ { type: "text" as const, text: `Unknown tool: ${request.params.name}`, }, ], }; }); return server; } async function main() { const apiKey = process.env.GEMINI_API_KEY; if (!apiKey) { console.error("Error: GEMINI_API_KEY environment variable is required"); console.error("Set it with: export GEMINI_API_KEY=your_key_here"); process.exit(1); } const server = createServer(apiKey); const transport = new StdioServerTransport(); await server.connect(transport); console.error("Nano Banana Pro MCP server started"); } // Only run main when executed directly (not when imported for testing) const isMainModule = import.meta.url === `file://${process.argv[1]}` || process.argv[1]?.endsWith("nano-banana-pro-mcp"); if (isMainModule) { main().catch((error) => { console.error("Fatal error:", error); process.exit(1); }); }