Vercel AI SDK

The @cloro-dev/ai-sdk npm package wraps cloro’s sync monitor endpoints as Vercel AI SDK tools. Each tool is a factory that returns an AI SDK tool() with a Zod input schema and a typed output. The source lives in the cloro-ai-sdk repository.

Prerequisites

ai 6 or 7, and zod 3.25.76 or later
A cloro API key from the dashboard (see Authentication)

Install the package

npm install @cloro-dev/ai-sdk ai zod

The tools read CLORO_API_KEY from the environment when they run. You can also pass it to each tool: webSearch({ apiKey: "YOUR_API_KEY" }).

Use with an agent

import { generateText, isStepCount } from "ai";
import { webSearch } from "@cloro-dev/ai-sdk";

const { text } = await generateText({
  model: "openai/gpt-5-mini",
  prompt: "What is the boiling point of water at 2,000 meters?",
  tools: {
    webSearch: webSearch(),
  },
  stopWhen: isStepCount(3),
});

With ai 6, use stepCountIs in place of isStepCount.

Available tools

Tool	Endpoint	The model sends	The tool returns
`webSearch`	`POST /v1/monitor/google`	`query`	The AI Overview text and its sources, and the organic results with title, link and snippet
`aiAnswer`	`POST /v1/monitor/chatgpt`, `gemini`, `aimode` or `bing/aimode`	`prompt`	The engine’s answer and the sources it cites

Both tools accept apiKey and country (ISO 3166-1 alpha-2, defaults to "US"). webSearch also takes maxResults (1 to 10) and includeAiOverview (defaults to true; false saves 2 credits). aiAnswer takes engine: "chatgpt" (default), "gemini", "google-ai-mode" or "bing-ai-mode". To give the model more than one engine, make one tool for each engine. Every call is a sync request, so it costs the provider rate plus the sync surcharge: 7 credits for webSearch with the AI Overview.

Introduction

Guides

Integrations

Vercel AI SDK

Prerequisites

Install the package

Use with an agent

Available tools

​Prerequisites

​Install the package

​Use with an agent

​Available tools

Prerequisites

Install the package

Use with an agent

Available tools