Uploads

You can easily upload files to websites using Playwright, Puppeteer, or Selenium. The approach varies depending on your framework:

Playwright

Direct Upload

For Playwright, you can upload files directly from your local path. After creating and connecting to a session, follow these steps:

1. Make sure your file is available where you’re running your Playwright code
2. Use the setInputFiles method to upload the file
3. The file path should be relative to your current working directory

Node.js

Playwright

import { chromium } from "playwright-core";
import { Browserbase } from "@browserbasehq/sdk";

(async () => {
  const bb = new Browserbase({ apiKey: process.env.BROWSERBASE_API_KEY! });
  const session = await bb.sessions.create();

  const browser = await chromium.connectOverCDP(session.connectUrl);
  const defaultContext = browser.contexts()[0];
  const page = defaultContext.pages()[0];

  await page.goto("https://browser-tests-alpha.vercel.app/api/upload-test");

  const fileInput = page.locator("#fileUpload");
  // logo.png is available relative to the current working directory
  await fileInput.setInputFiles("logo.png");
})().catch((error) => console.error(error));

Python

Playwright

from playwright.sync_api import sync_playwright
from browserbase import Browserbase
import os

bb = Browserbase(api_key=os.environ["BROWSERBASE_API_KEY"])
session = bb.sessions.create()

with sync_playwright() as playwright:
    browser = playwright.chromium.connect_over_cdp(session.connect_url)
    context = browser.contexts[0]
    page = context.pages[0]

    page.goto("https://browser-tests-alpha.vercel.app/api/upload-test")

    file_input = page.locator("#fileUpload")
    file_input.set_input_files("logo.png")

Large Files Upload

For larger files, you can use the Session Uploads API. Follow this example:

Node.js

Playwright

// ⭐️⭐️⭐️ IMPORTANT: Set your file name below
const fileName = "YOUR_FILE_NAME.EXAMPLE";
// ⭐️⭐️⭐️

import { chromium } from "playwright-core";
import { Browserbase } from "@browserbasehq/sdk";
import * as fs from "fs";

const apiKey = process.env.BROWSERBASE_API_KEY!;

async function upload() {
  // 1. Initialize Browserbase Client
  console.log("✨ Initializing Browserbase client");
  const bb = new Browserbase({ apiKey });

  // 2. Create Browser Session
  console.log("🚀 Creating new browser session");
  const session = await bb.sessions.create();

  // 3. Upload file via the Uploads API
  console.log("⬆️ Uploading file");
  try {
    const fileStream = fs.createReadStream(fileName);
    const result = await bb.sessions.uploads.create(session.id, {
      file: fileStream,
    });
    console.log(`✅ Upload successful: ${JSON.stringify(result)}`);
  } catch (error) {
    console.error(`❌ Upload failed... exiting: ${error}`);
    return;
  }

  // 4. Connect to the Session
  console.log("🔗 Connecting to browser session");
  const browser = await chromium.connectOverCDP(session.connectUrl);
  const defaultContext = browser.contexts()[0];
  const page = defaultContext.pages()[0];

  // 5. Get Live View link for remote debugging
  const liveViews = await bb.sessions.debug(session.id);
  console.log("🔍 Live View link:", liveViews.debuggerUrl);

  // 6. Use the Browser
  console.log("🌐 Navigating to page: upload-test");
  await page.goto("https://browser-tests-alpha.vercel.app/api/upload-test", {
    waitUntil: "domcontentloaded",
  });

  // Set up CDP client for additional controls
  const cdpSession = await defaultContext.newCDPSession(page);
  const root = await cdpSession.send("DOM.getDocument");

  // Then find our input element
  const inputNode = await cdpSession.send("DOM.querySelector", {
    nodeId: root.root.nodeId,
    selector: "#fileUpload",
  });

  // Use DOM.setFileInputFiles CDP command
  const remoteFilePath = `/tmp/.uploads/${fileName}`;
  await cdpSession.send("DOM.setFileInputFiles", {
    files: [remoteFilePath],
    nodeId: inputNode.nodeId,
  });

  console.log("⌛ Waiting for 60 seconds: allow time for 1) file upload and 2) to see the file upload...");
  await new Promise((resolve) => setTimeout(resolve, 60000));

  // 7. Cleanup
  console.log("👋 Closing browser session");
  await page.close();
  await browser.close();

  // 8. Session Recording Link
  console.log("\n" + "─".repeat(60));
  console.log(`
🎥  Your session dashboard is ready
  https://www.browserbase.com/sessions/${session.id}`);
}

// Execute the main function
upload().catch((error) => console.error(error));

Python

Playwright

# ⭐️⭐️⭐️ IMPORTANT: Set your file name below
file_name = "YOUR_FILE_NAME.EXAMPLE"
# ⭐️⭐️⭐️

import os
from browserbase import Browserbase
from playwright.sync_api import sync_playwright
import time

api_key = os.environ.get("BROWSERBASE_API_KEY")

def upload(playwright):
  # 1. Initialize Browserbase Client
  # ───────────────────────────────
  print("✨ Initializing Browserbase client")
  bb = Browserbase(api_key=api_key)

  # 2. Create Browser Session
  # ───────────────────────────
  print("🚀 Creating new browser session")
  session = bb.sessions.create()

  # 3. Upload file via the Uploads API: https://docs.browserbase.com/reference/api/create-session-uploads
  # ───────────────────────
  print("⬆️ Uploading file")
  try:
    with open(file_name, 'rb') as file:
      result = bb.sessions.uploads.create(
        id=session.id,
        file=file
    )
      print(f"✅ Upload successful: {result}")
  except Exception as e:
    print(f"❌ Upload failed... exiting: {str(e)}")
    return

  # 4. Connect to the Session
  # ───────────────────────
  print("🔗 Connecting to browser session")
  browser = playwright.chromium.connect_over_cdp(session.connect_url)
  default_context = browser.contexts[0]
  page = default_context.pages[0]

  # 5. Get Live View link for remote debugging
  # ───────────────────────────────
  live_views = bb.sessions.debug(session.id)
  print("🔍 Live View link:", live_views.debugger_url)

  # 6. Use the Browser
  # ─────────────────────────
  print("🌐 Navigating to page: upload-test")
  page.goto("https://browser-tests-alpha.vercel.app/api/upload-test", wait_until="domcontentloaded")

  # Set up CDP client for additional controls
  cdp_session = default_context.new_cdp_session(page)

  root = cdp_session.send('DOM.getDocument')

  # Then find our input element
  input_node = cdp_session.send('DOM.querySelector', {
    'nodeId': root['root']['nodeId'],
    'selector': '#fileUpload'
  })

  # Use DOM.setFileInputFiles CDP command with the found nodeId and set the remote file path
  remote_file_path = f"/tmp/.uploads/{file_name}";
  cdp_session.send('DOM.setFileInputFiles', {
    'files': [remote_file_path],
    'nodeId': input_node['nodeId']
  })

  print("⌛ Waiting for 60 seconds: allow time for 1) file upload and 2) to see the file upload...")
  time.sleep(60)

  # 7. Cleanup
  # ───────────────────────
  print("👋 Closing browser session")
  page.close()
  browser.close()

  # 8. Session Recording Link
  # ────────────────────────────
  print("\n" + "─" * 60)
  print(f"""
🎥  Your session dashboard is ready
  https://www.browserbase.com/sessions/{session.id}""")

# Execute the main function
if __name__ == "__main__":
    with sync_playwright() as playwright:
        upload(playwright)

Manual Upload through Live View

When using the Live View to interact with a browser session, clicking a file input triggers the browser’s native file picker. However, since the browser runs remotely on Browserbase infrastructure, it cannot access files on your local machine.

To enable manual uploads through the live view, one workaround you can implement is to intercept file chooser events and handle uploads programmatically:

1. Detect file input clicks using Playwright’s page.on('filechooser') event
2. Show your own file picker (e.g., a native OS picker or web UI) to select a local file
3. Upload the file via the Session Uploads API
4. Attach the file to the input element using CDP’s DOM.setFileInputFiles

See the following example script for a complete implementation.

Session Uploads API

Learn more about the available params and response fields