Vercel · AI SDK · Connect

Bring live web sources into your Vercel apps

Give your AI SDK agent search, news, and page reading tools. Use Vercel Connect when your application needs user-authorized Search1API credentials at runtime.

Choose your Vercel integration

Start with the tools package for AI SDK agents, or the connector for credentials managed through Vercel.

Vercel AI SDK tools

Ready-to-use tools for your agent.

Search the web, find news, and read page content with @search1api/ai-sdk.

Works with generateText and streamText in AI SDK apps, including Next.js.

Install the npm package and set your Search1API API key on the server.

Vercel Connect

Retrieve Search1API credentials at runtime.

Connect a Search1API account and request access with @vercel/connect.

For Vercel projects with access scoped to the project and environment.

Create the Search1API connector, authorize the account, then request a token for the user.

Search, read, and cite current sources

Live web research

Find current documentation, news, and public sources while your application runs.

Agent-ready tools

The AI SDK package handles tool schemas and execution so your agent can search and read pages.

Runtime credentials

Vercel Connect retrieves credentials on demand for the authorized user.

Build with current information

Add a search-and-read workflow to a Next.js chat app.

Research recent news and return answers with source links.

Use an authorized Search1API account from your Vercel application.

Get started with either integration

Use the AI SDK package for tool calling. Run the Connect commands from the Vercel project that will use the connector. Follow the Vercel guide for authorization and runtime code.

AI SDK tools: install and configure — bash
npm install @search1api/ai-sdk ai zod
export SEARCH1API_API_KEY="your-api-key"

AI SDK tools: search and read with your agent — ts
import { gateway, generateText, stepCountIs } from 'ai';
import { search1apiTools } from '@search1api/ai-sdk';
const result = await generateText({
  model: gateway('openai/gpt-5-mini'),
  tools: search1apiTools({ only: ['search', 'crawl'] }),
  stopWhen: stepCountIs(5),
  prompt: 'Find the AI SDK tool calling docs, read them, and cite the URLs you used.',
});
console.log(result.text);

Vercel Connect: create your connector — bash
npm install @vercel/connect
vercel link
vercel connect create search1api --name acme-search1api
vercel env pull

Before you run

Keep Search1API API keys and Connect tokens in server-side code.

The AI SDK package uses your Search1API API key. Your model provider requires its own credentials.

Vercel Connect requires a linked project, an enabled environment, and authorization for the requested user.

Search1API requests use the connected account’s credits. See the guide for billing details.

FAQ

Which integration should I choose?

Use @search1api/ai-sdk when you want ready-made AI SDK tools. Use Vercel Connect when you want Vercel to manage runtime access to an authorized Search1API account.

Does the AI SDK package require Vercel hosting?

No. It works in compatible server-side JavaScript runtimes wherever you run your AI SDK application.

Does Connect provide AI SDK tools?

Connect supplies runtime credentials. The @search1api/ai-sdk package provides the search, news, and crawl tool definitions and execution.