/**
* generate_image MCP Tool
* Generate images using xAI's Grok Imagine model
*/
import { z } from "zod";
import { getXAIClient } from "../xai-client.js";
export const generateImageSchema = z.object({
prompt: z
.string()
.describe("Text description of the image to generate"),
n: z
.number()
.min(1)
.max(10)
.optional()
.default(1)
.describe("Number of images to generate (1-10)"),
model: z
.string()
.optional()
.default("grok-2-image")
.describe("Image generation model (default: grok-2-image)"),
aspect_ratio: z
.string()
.optional()
.describe("Aspect ratio (e.g., '16:9', '1:1', '4:3')"),
response_format: z
.enum(["url", "b64_json"])
.optional()
.default("url")
.describe("Response format: 'url' for hosted URL or 'b64_json' for base64"),
});
export type GenerateImageInput = z.infer<typeof generateImageSchema>;
export const generateImageTool = {
name: "generate_image",
description:
"Generate images from text descriptions using xAI's Grok Imagine model. Returns image URLs or base64 data.",
inputSchema: {
type: "object" as const,
properties: {
prompt: {
type: "string",
description: "Text description of the image to generate",
},
n: {
type: "number",
description: "Number of images to generate (1-10)",
default: 1,
},
model: {
type: "string",
description: "Image generation model (default: grok-2-image)",
default: "grok-2-image",
},
aspect_ratio: {
type: "string",
description: "Aspect ratio (e.g., '16:9', '1:1', '4:3')",
},
response_format: {
type: "string",
enum: ["url", "b64_json"],
description: "Response format: 'url' or 'b64_json'",
default: "url",
},
},
required: ["prompt"],
},
};
export async function handleGenerateImage(
input: GenerateImageInput
): Promise<string> {
const client = getXAIClient();
const validated = generateImageSchema.parse(input);
const response = await client.generateImage({
model: validated.model,
prompt: validated.prompt,
n: validated.n,
response_format: validated.response_format,
aspect_ratio: validated.aspect_ratio,
});
if (validated.response_format === "b64_json") {
const images = response.data.map((img, i) => ({
index: i + 1,
base64: img.b64_json?.substring(0, 100) + "...", // Truncate for display
revised_prompt: img.revised_prompt,
}));
return JSON.stringify({ success: true, images, note: "Base64 data truncated for display" }, null, 2);
}
const images = response.data.map((img, i) => ({
index: i + 1,
url: img.url,
revised_prompt: img.revised_prompt,
}));
return JSON.stringify({ success: true, images }, null, 2);
}
