Searchclaw

Instructions

Implementation Reference

• src/index.ts:192-193 (handler)

  The handler function for the 'pipeline' tool that executes the search + extract logic. It calls apiPost('/pipeline', ...) with the query, schema, max_results, and extract_from parameters, then formats the result using jsonResult().

  async ({ query, schema, max_results, extract_from }) =>
    jsonResult(await apiPost("/pipeline", { query, schema, max_results, extract_from }))

• src/index.ts:186-191 (schema)

  Input schema definition for the 'pipeline' tool using Zod. Defines 4 parameters: query (string, required), schema (record, required), max_results (number, optional, default 10), and extract_from (number, optional, default 5).

  {
    query: z.string().describe("Search query to find relevant pages"),
    schema: z.record(z.unknown()).describe("JSON schema defining data to extract from results"),
    max_results: z.number().optional().default(10).describe("Max search results (default: 10)"),
    extract_from: z.number().optional().default(5).describe("Number of top results to extract from (default: 5)"),
  },

• src/index.ts:183-194 (registration)

  Complete registration of the 'pipeline' tool with the MCP server using server.tool(). Includes tool name, description, input schema, and handler function.

  server.tool(
    "pipeline",
    "Search + extract in one call. The killer feature — find pages via search, then extract structured data from top results. Costs 3+ credits.",
    {
      query: z.string().describe("Search query to find relevant pages"),
      schema: z.record(z.unknown()).describe("JSON schema defining data to extract from results"),
      max_results: z.number().optional().default(10).describe("Max search results (default: 10)"),
      extract_from: z.number().optional().default(5).describe("Number of top results to extract from (default: 5)"),
    },
    async ({ query, schema, max_results, extract_from }) =>
      jsonResult(await apiPost("/pipeline", { query, schema, max_results, extract_from }))
  );

• src/index.ts:41-59 (helper)

  Helper function apiPost() that makes HTTP POST requests to the SearchClaw API with timeout handling, error handling, and JSON response parsing. Used by the pipeline tool handler.

  async function apiPost(path: string, body: Record<string, unknown>) {
    const controller = new AbortController();
    const timeout = setTimeout(() => controller.abort(), 30000);
    try {
      const response = await fetch(`${API_BASE}${path}`, {
        method: "POST",
        headers: { ...headers, "Content-Type": "application/json" },
        body: JSON.stringify(body),
        signal: controller.signal,
      });
      if (!response.ok) {
        const text = await response.text();
        throw new Error(`SearchClaw API error ${response.status}: ${text}`);
      }
      return response.json();
    } finally {
      clearTimeout(timeout);
    }
  }

• src/index.ts:61-63 (helper)

  Helper function jsonResult() that formats API responses into the MCP tool result format with proper content type and JSON stringification.

  function jsonResult(data: unknown) {
    return { content: [{ type: "text" as const, text: JSON.stringify(data, null, 2) }] };
  }