/**
* analyze_image MCP Tool
* Analyze images using xAI's vision-capable Grok models
* Based on: https://docs.x.ai/docs/guides/chat-completions
*/
import { z } from "zod";
import { getXAIClient } from "../xai-client.js";
export const visionSchema = z.object({
image_url: z
.string()
.describe("URL of the image to analyze (JPEG, PNG, max 20MB) or base64 data URL"),
prompt: z
.string()
.optional()
.default("Describe this image in detail.")
.describe("Question or instruction about the image"),
detail: z
.enum(["low", "high", "auto"])
.optional()
.default("auto")
.describe("Image detail level: 'low' (faster, fewer tokens), 'high' (detailed), 'auto' (model decides)"),
model: z
.string()
.optional()
.default("grok-2-vision-1212")
.describe("Vision-capable model (grok-2-vision-1212)"),
});
export type VisionInput = z.infer<typeof visionSchema>;
export const visionTool = {
name: "analyze_image",
description:
"Analyze images using xAI's vision-capable Grok models. " +
"Describe images, extract text (OCR), answer questions about visual content, " +
"or identify objects and scenes. Supports JPEG and PNG up to 20MB.",
inputSchema: {
type: "object" as const,
properties: {
image_url: {
type: "string",
description: "URL of the image to analyze (JPEG, PNG, max 20MB) or base64 data URL",
},
prompt: {
type: "string",
description: "Question or instruction about the image",
default: "Describe this image in detail.",
},
detail: {
type: "string",
enum: ["low", "high", "auto"],
description: "Image detail level for analysis",
default: "auto",
},
model: {
type: "string",
description: "Vision-capable model (grok-2-vision-1212)",
default: "grok-2-vision-1212",
},
},
required: ["image_url"],
},
};
export async function handleVision(input: VisionInput): Promise<string> {
const client = getXAIClient();
const validated = visionSchema.parse(input);
const analysis = await client.analyzeImage(
validated.image_url,
validated.prompt,
validated.model,
validated.detail
);
return JSON.stringify(
{
success: true,
analysis,
image_url: validated.image_url,
prompt: validated.prompt,
model: validated.model,
detail: validated.detail,
},
null,
2
);
}
