batch_fetch

Instructions

Implementation Reference

• src/index.ts:162-230 (handler)

  The 'batch_fetch' tool handler: calls the /batch endpoint, processes results, and formats markdown output.

  server.tool(
    "batch_fetch",
    "Fetch multiple URLs in parallel and convert them all to clean Markdown. Process up to 10 URLs in a single call.",
    {
      urls: z
        .array(z.string().url())
        .min(1)
        .max(10)
        .describe("Array of URLs to fetch (1-10)"),
      model: z
        .string()
        .optional()
        .describe("AI model ID for cost tracking"),
    },
    async (params) => {
      const apiKey = getApiKey();
  
      const response = await fetch(`${BASE_URL}/batch`, {
        method: "POST",
        headers: {
          Authorization: `Bearer ${apiKey}`,
          "Content-Type": "application/json",
        },
        body: JSON.stringify({
          urls: params.urls,
          model: params.model,
        }),
      });
  
      if (!response.ok) {
        const body = await response.text();
        let message: string;
        try {
          message = JSON.parse(body).error;
        } catch {
          message = body;
        }
        throw new Error(`StripFeed API error ${response.status}: ${message}`);
      }
  
      const data = await response.json();
      const results = data.results as Array<{
        url: string;
        title: string;
        markdown: string;
        tokens: number;
        originalTokens: number;
        savingsPercent: number;
        status: number;
        error?: string;
      }>;
  
      const sections = results.map((r) => {
        if (r.status !== 200) {
          return `## ${r.url}\n\nError: ${r.error ?? `Status ${r.status}`}`;
        }
        const saved = `${r.tokens.toLocaleString()} tokens (saved ${r.savingsPercent}% from ${r.originalTokens.toLocaleString()})`;
        return `## ${r.title || r.url}\n\nSource: ${r.url} | ${saved}\n\n${r.markdown}`;
      });
  
      const summary = `Fetched ${data.success}/${data.total} URLs successfully.`;
  
      return {
        content: [
          { type: "text" as const, text: `${summary}\n\n---\n\n${sections.join("\n\n---\n\n")}` },
        ],
      };
    }
  );

• src/index.ts:162-175 (schema)

  The input schema for batch_fetch: takes an array of URLs (1-10) and an optional model string.

  server.tool(
    "batch_fetch",
    "Fetch multiple URLs in parallel and convert them all to clean Markdown. Process up to 10 URLs in a single call.",
    {
      urls: z
        .array(z.string().url())
        .min(1)
        .max(10)
        .describe("Array of URLs to fetch (1-10)"),
      model: z
        .string()
        .optional()
        .describe("AI model ID for cost tracking"),
    },

• src/index.ts:162-163 (registration)

  Registration of the 'batch_fetch' tool on the MCP server via server.tool().

  server.tool(
    "batch_fetch",