Sidearm MCP Server

Instructions

Implementation Reference

• src/tools/detect-ai.ts:5-70 (handler)

  Main implementation of detect_ai tool - defines the tool with Zod schema for input validation (media_url, media, text, mime, tags) and the handler function that posts to /api/v1/detect/ai endpoint, returning a job_id for polling via check_job

  export function register(server: McpServer, api: ApiClient): void {
    server.tool(
      "detect_ai",
      "Detect whether media content was generated by AI. Supports images, video, audio, " +
        "and text/PDF. Runs multiple specialized detection models in parallel for the given " +
        "media type. Returns a job_id — use check_job to poll for results.",
      {
        media_url: z
          .string()
          .url()
          .optional()
          .describe("Public URL of the media to analyze"),
        media: z
          .string()
          .optional()
          .describe("Base64-encoded media content to analyze"),
        text: z
          .string()
          .optional()
          .describe("Plain text content to analyze for AI generation"),
        mime: z
          .string()
          .optional()
          .describe("MIME type of the media (e.g. image/png, audio/wav, text/plain)"),
        tags: z
          .array(z.string())
          .optional()
          .describe("Tags for organizing and filtering"),
      },
      async (params) => {
        try {
          const body: Record<string, unknown> = {};
          if (params.media_url) body.media_url = params.media_url;
          if (params.media) body.media = params.media;
          if (params.text) body.text = params.text;
          if (params.mime) body.mime = params.mime;
          if (params.tags) body.tags = params.tags;
  
          const result = await api.post("/api/v1/detect/ai", body);
          const res = result as { job_id: string };
  
          return {
            content: [
              {
                type: "text" as const,
                text:
                  `AI detection job created.\n\n` +
                  `Job ID: ${res.job_id}\n\n` +
                  `Use check_job with this job_id to poll for results.`,
              },
            ],
          };
        } catch (err) {
          return {
            content: [
              {
                type: "text" as const,
                text: `Error: ${err instanceof Error ? err.message : String(err)}`,
              },
            ],
            isError: true as const,
          };
        }
      },
    );
  }

• src/tools/detect-ai.ts:12-33 (schema)

  Zod input schema definition for detect_ai tool parameters including optional media_url, media, text, mime type, and tags fields

    media_url: z
      .string()
      .url()
      .optional()
      .describe("Public URL of the media to analyze"),
    media: z
      .string()
      .optional()
      .describe("Base64-encoded media content to analyze"),
    text: z
      .string()
      .optional()
      .describe("Plain text content to analyze for AI generation"),
    mime: z
      .string()
      .optional()
      .describe("MIME type of the media (e.g. image/png, audio/wav, text/plain)"),
    tags: z
      .array(z.string())
      .optional()
      .describe("Tags for organizing and filtering"),
  },

• src/index.ts:12-12 (registration)

  Import statement for detect_ai register function

  import { register as detectAi } from "./tools/detect-ai.js";

• src/index.ts:54-54 (registration)

  Registration of detect_ai tool with the MCP server instance and API client

  detectAi(server, api);

• src/api.ts:25-31 (helper)

  ApiClient.post method used by detect_ai handler to make API requests to the SDRM API endpoint

  async post<T = unknown>(path: string, body: unknown): Promise<T> {
    return this.request<T>(new URL(`${this.baseUrl}${path}`), {
      method: "POST",
      headers: { "Content-Type": "application/json" },
      body: JSON.stringify(body),
    });
  }