Vercel AI SDK tool with Stophy | Stophy

Give any AI SDK model live web search. The model asks for a search, execute calls Stophy, and the model answers from the results.

The search tool

Save this as stophy.ts. It calls Google search and returns the top 5 results as JSON. It works with no key. Set STOPHY_API_KEY to lift the free limit.

stophy.ts

export async function stophySearch(query: string): Promise<string> {
  const headers: Record<string, string> = { "content-type": "application/json" };
  if (process.env.STOPHY_API_KEY) headers.authorization = `Bearer ${process.env.STOPHY_API_KEY}`;
  const response = await fetch("https://api.stophy.dev/v1/google/search", {
    method: "POST",
    headers,
    body: JSON.stringify({ query }),
  });
  const body = await response.json();
  if (!body.success) throw new Error(body.error.message);
  return JSON.stringify(body.data.results.slice(0, 5));
}

Give it to the AI SDK

npm install ai @ai-sdk/openai zod

Set OPENAI_API_KEY in your environment. Wrap the search in tool(). stopWhen lets the model search, read the results and answer in up to five steps. Any AI SDK model works, so swap openai(...) for another provider.

import { openai } from "@ai-sdk/openai";
import { generateText, stepCountIs, tool } from "ai";
import { z } from "zod";
import { stophySearch } from "./stophy";

const { text } = await generateText({
  model: openai("gpt-5-mini"),
  prompt: "What changed in the latest Bun release?",
  tools: {
    webSearch: tool({
      description: "Search the web. Returns the top results as JSON.",
      inputSchema: z.object({ query: z.string().describe("What to search for") }),
      execute: async ({ query }) => stophySearch(query),
    }),
  },
  stopWhen: stepCountIs(5),
});

console.log(text);