Mock Video Injection on Web Automation (Selenium/Chromium)

Inject a custom video file as a fake camera feed in Chromium-based Selenium tests on TestMu AI. This is useful for testing WebRTC, video conferencing, KYC or identity verification, and any flow that calls getUserMedia().

How It Works

1. Upload your .mjpeg or .y4m video file to TestMu AI using the web automation user-files API.
2. Pre-load the file onto the test VM using the lambda:userFiles capability.
3. Pass Chrome flags to use the uploaded file as a fake camera device.

Chrome treats the file as a looping camera feed, so any getUserMedia({ video: true }) call receives frames from your file instead of a real camera.

Prerequisites

• A TestMu AI account with Web Automation access
• A .mjpeg video file such as sample_640x360.mjpeg
  • Recommended resolution: 640x360 or 1280x720
  • Chrome also supports .y4m (raw YUV4MPEG2) format
• TestMu AI credentials available as LT_USERNAME and LT_ACCESS_KEY

Step 1: Upload the Video File

Upload your .mjpeg file using the web automation user-files API.

cURL

curl -X POST \
  "https://api.lambdatest.com/automation/api/v1/user-files" \
  -u "$LT_USERNAME:$LT_ACCESS_KEY" \
  -F "files=@/path/to/sample_640x360.mjpeg"

Response

{
  "status": "success",
  "data": [
    {
      "error": "",
      "message": "File have been uploaded successfully to our lambda storage",
      "key": "sample_640x360.mjpeg"
    }
  ]
}

note

Limits: maximum 20 MB per upload and maximum 150 files per organization.

The /mfs/v1.0/media/upload endpoint is for mobile and app automation. For web automation, use /automation/api/v1/user-files.

Step 2: Configure Selenium Capabilities

Java (Selenium 4, W3C)

import org.openqa.selenium.chrome.ChromeOptions;
import org.openqa.selenium.remote.RemoteWebDriver;
import java.net.URL;
import java.util.*;

ChromeOptions options = new ChromeOptions();

options.addArguments(
    "--use-fake-ui-for-media-stream",
    "--use-fake-device-for-media-stream",
    "--use-file-for-fake-video-capture=/home/ltuser/Downloads/sample_640x360.mjpeg"
);

HashSet<String> userFiles = new HashSet<>();
userFiles.add("sample_640x360.mjpeg");

HashMap<String, Object> ltOptions = new HashMap<>();
ltOptions.put("platform", "Linux");
ltOptions.put("build", "Mock Video Injection Test");
ltOptions.put("name", "Fake Camera Feed Test");
ltOptions.put("video", true);
ltOptions.put("w3c", true);

options.setCapability("LT:Options", ltOptions);
options.setCapability("lambda:userFiles", userFiles);

RemoteWebDriver driver = new RemoteWebDriver(
    new URL("https://" + LT_USERNAME + ":" + LT_ACCESS_KEY + "@hub.lambdatest.com/wd/hub"),
    options
);

Python

from selenium import webdriver

options = webdriver.ChromeOptions()
options.add_argument("--use-fake-ui-for-media-stream")
options.add_argument("--use-fake-device-for-media-stream")
options.add_argument("--use-file-for-fake-video-capture=/home/ltuser/Downloads/sample_640x360.mjpeg")

lt_options = {
    "platform": "Linux",
    "build": "Mock Video Injection Test",
    "name": "Fake Camera Feed Test",
    "video": True,
    "w3c": True,
}

options.set_capability("LT:Options", lt_options)
options.set_capability("lambda:userFiles", ["sample_640x360.mjpeg"])

driver = webdriver.Remote(
    command_executor=f"https://{LT_USERNAME}:{LT_ACCESS_KEY}@hub.lambdatest.com/wd/hub",
    options=options,
)

JavaScript (WebDriverIO or raw WebDriver)

const capabilities = {
  browserName: "Chrome",
  browserVersion: "latest",
  platformName: "Linux",
  "LT:Options": {
    platform: "Linux",
    build: "Mock Video Injection Test",
    name: "Fake Camera Feed Test",
    video: true,
    w3c: true,
  },
  "lambda:userFiles": ["sample_640x360.mjpeg"],
  "goog:chromeOptions": {
    args: [
      "--use-fake-ui-for-media-stream",
      "--use-fake-device-for-media-stream",
      "--use-file-for-fake-video-capture=/home/ltuser/Downloads/sample_640x360.mjpeg",
      "--no-sandbox",
      "--disable-gpu",
    ],
  },
};

Step 3: Verify the Video Feed in Your Test

After the session starts, navigate to a page that requests camera access and confirm the fake stream is being used.

driver.get("https://www.lambdatest.com/selenium-playground/webrtc-video");

// Or open your application's camera page
driver.get("https://your-app.com/video-call");

Boolean isPlaying = (Boolean) driver.executeScript(
    "const video = document.querySelector('video');" +
    "return video && !video.paused && video.readyState >= 2;"
);

File Paths by OS

The lambda:userFiles capability places files in these directories:

Platform	File Path
Linux	/home/ltuser/Downloads/sample_640x360.mjpeg
Windows	C:\\Users\\ltuser\\Downloads\\sample_640x360.mjpeg
macOS	/Users/ltuser/Downloads/sample_640x360.mjpeg

Adjust the --use-file-for-fake-video-capture path to match your target platform.

Chrome Flags Reference

Flag	Purpose
--use-fake-ui-for-media-stream	Auto-grants getUserMedia() permission without a popup
--use-fake-device-for-media-stream	Replaces real camera and microphone hardware with fake devices
--use-file-for-fake-video-capture=<path>	Uses the specified file as the fake camera feed
--use-file-for-fake-audio-capture=<path>	Uses the specified file as fake microphone input (.wav format)

Troubleshooting

Issue	Solution
Black or empty video feed	Verify the file name in lambda:userFiles matches exactly, including case
Permission popup still appears	Confirm --use-fake-ui-for-media-stream is set
File not found on VM	Check the OS-specific path; if needed, use driver.executeScript("lambda-files-download=sample_640x360.mjpeg") as a fallback
Video not looping	Chrome loops .mjpeg files automatically; .y4m files also loop by default
Wrong resolution	Use a file resolution that matches your app expectations; 640x360 is a safe default
Windows path errors	Use escaped backslashes: C:\\Users\\ltuser\\Downloads\\sample_640x360.mjpeg
macOS video does not render	This is a known Chrome limitation; use Linux when visual verification is required