@browserbasehq/sdk

Documentation  ·  Playground

Browserbase JS SDK

Browserbase is the best developer platform to reliably run, manage, and monitor headless browsers. Leverage Browserbase to power your automation, test suites, and LLM data retrievals.

Installation and setup

1. Install the SDK:

npm install @browserbasehq/sdk

2. Get your Browserbase API Key and Project ID:

• Create an account or log in to Browserbase
• Copy your API Key and Project ID from your Settings page

Usage

import { Browserbase } from '@browserbasehq/sdk'

// Init the SDK
const browserbase = new Browserbase()

// Load a webpage
const rawHtml = await browserbase.load('https://www.browserbase.com')

// Load multiple webpages (returns iterator)
const rawHtmls = browserbase.loadURLs([
  'https://www.browserbase.com',
  'https://docs.browserbase.com',
])

for await (let rawHtml of rawHtmls) {
  // ...
}

// Text-only mode
const text = await browserbase.load('https://www.browserbase.com', {
  textContent: true,
})

// Screenshot (returns bytes)
const result = await browserbase.screenshot('https://www.browserbase.com', {
  textContent: true,
})

Vercel AI SDK Integration

Install the additional dependencies:

npm install ai openai zod

import OpenAI from 'openai'
import { Browserbase, BrowserbaseAISDK } from '@browserbasehq/sdk'
import {
  OpenAIStream,
  StreamingTextResponse,
  generateText,
} from 'ai'

// Create new OpenAI client
const openai = new OpenAI({
  apiKey: process.env.OPENAI_API_KEY,
})

// Init the Browserbase SDK
const browserbase = new Browserbase()

// Init the tool
const browserTool = BrowserbaseAISDK(browserbase, { textContent: true })

// Load completions
const result = await generateText({
  model: openai.chat('gpt-3.5-turbo'),
  tools: {
    browserTool,
  },
  prompt: 'What is the weather in San Francisco?',
})

Further reading

• See how to leverage the Session Live View for faster development
• Sessions API Reference