playwright Pod¶

Standard OCI container - works with Docker, Podman, Kubernetes, Apptainer.

The playwright pod provides a complete browser automation environment with Playwright, Google Chrome, a Wayland compositor (Sway), and VNC access for visual debugging.

Overview¶

Attribute	Value
Image	`ghcr.io/atrawog/bazzite-ai-pod-playwright:stable`
Size	~5GB
GPU	NVIDIA (optional, for GPU acceleration)
Inherits	pod-nvidia
Port	5900 (VNC)

Quick Start¶

Docker / PodmanKubernetesHPC (Apptainer)Bazzite AI OS

# With VNC access (connect to localhost:5900)
docker run -it --rm -p 5900:5900 -v $(pwd):/workspace \
  ghcr.io/atrawog/bazzite-ai-pod-playwright:stable

# With GPU acceleration
docker run -it --rm --gpus all --privileged -p 5900:5900 -v $(pwd):/workspace \
  ghcr.io/atrawog/bazzite-ai-pod-playwright:stable

# Headless only (no VNC)
docker run -it --rm -v $(pwd):/workspace \
  ghcr.io/atrawog/bazzite-ai-pod-playwright:stable bash

apiVersion: v1
kind: Pod
metadata:
  name: playwright-test
spec:
  containers:
  - name: playwright
    image: ghcr.io/atrawog/bazzite-ai-pod-playwright:stable
    ports:
    - containerPort: 5900
      name: vnc
    volumeMounts:
    - name: workspace
      mountPath: /workspace
  volumes:
  - name: workspace
    persistentVolumeClaim:
      claimName: test-workspace

apptainer pull docker://ghcr.io/atrawog/bazzite-ai-pod-playwright:stable
apptainer exec bazzite-ai-pod-playwright_stable.sif bash

apptainer pull docker://ghcr.io/atrawog/bazzite-ai-pod-playwright:stable
apptainer shell bazzite-ai-pod-playwright_stable.sif

What's Included¶

Browser Automation¶

Component	Description
Playwright	Browser automation framework (Python)
Google Chrome	Full Chrome browser for testing
Playwright browsers	Pre-installed browser binaries

Display Environment¶

Component	Description
Sway	Wayland compositor (tiling window manager)
wayvnc	VNC server for remote access
s6-overlay	Process supervision (container init)

VR/Streaming Support¶

Component	Description
OpenVR	VR headset support
OpenXR	Cross-platform VR/AR API
OpenCV	Computer vision library
GStreamer	Media streaming framework

From nvidia Pod¶

CUDA Toolkit 13.0
cuDNN, TensorRT

From base Pod¶

Python, Node.js, VS Code, Docker CLI
kubectl, Helm, Claude Code

Usage¶

VNC Access¶

Connect with any VNC client:

Address: localhost:5900
Password: None (local development)

Recommended clients:

macOS: Built-in Screen Sharing (vnc://localhost:5900)
Linux: Remmina, TigerVNC
Windows: RealVNC, TightVNC

Running Playwright Tests¶

Inside the pod (via VNC or exec):

# Activate the environment
pixi shell --manifest-path /opt/pixi/pixi.toml

# Run tests
pytest tests/

# Or run directly
pixi run --manifest-path /opt/pixi/pixi.toml pytest tests/

Writing Playwright Tests¶

# tests/test_example.py
from playwright.sync_api import sync_playwright

def test_homepage():
    with sync_playwright() as p:
        browser = p.chromium.launch(headless=False)
        page = browser.new_page()
        page.goto("https://example.com")

        assert page.title() == "Example Domain"

        browser.close()

def test_screenshot():
    with sync_playwright() as p:
        browser = p.chromium.launch()
        page = browser.new_page()
        page.goto("https://example.com")

        # Screenshot saved to /workspace
        page.screenshot(path="/workspace/screenshot.png")

        browser.close()

Async Playwright¶

import asyncio
from playwright.async_api import async_playwright

async def main():
    async with async_playwright() as p:
        browser = await p.chromium.launch(headless=False)
        page = await browser.new_page()
        await page.goto("https://example.com")

        # Interact with the page
        await page.click("text=More information")

        await browser.close()

asyncio.run(main())

Common Workflows¶

Visual Test Debugging¶

Run pod with VNC
Connect VNC client
Run tests in headed mode (headless=False)
Watch browser interactions in real-time

Screenshot Testing¶

def test_visual_regression():
    with sync_playwright() as p:
        browser = p.chromium.launch()
        page = browser.new_page()
        page.goto("https://myapp.com")

        # Compare to baseline
        page.screenshot(path="/workspace/screenshots/current.png")

        browser.close()

Form Testing¶

def test_login_form():
    with sync_playwright() as p:
        browser = p.chromium.launch()
        page = browser.new_page()
        page.goto("https://myapp.com/login")

        # Fill form
        page.fill('input[name="email"]', 'test@example.com')
        page.fill('input[name="password"]', 'secret123')
        page.click('button[type="submit"]')

        # Verify redirect
        assert page.url == "https://myapp.com/dashboard"

        browser.close()

Environment Details¶

Display Configuration¶

Variable	Value
`XDG_RUNTIME_DIR`	`/run/user/1000`
`WAYLAND_DISPLAY`	`wayland-0`
`XDG_SESSION_TYPE`	`wayland`

Playwright Configuration¶

Variable	Value
`PLAYWRIGHT_BROWSERS_PATH`	`/home/jovian/.cache/ms-playwright`
`PLAYWRIGHT_PROJECT`	`/opt/pixi`

Troubleshooting¶

VNC Connection Refused¶

Ensure port 5900 is published: -p 5900:5900
Check if Sway started: docker logs <container>
Verify s6-overlay initialization

Browser Won't Launch¶

# Use explicit executable path
browser = p.chromium.launch(
    executable_path="/usr/bin/google-chrome-stable",
    headless=False
)

Display Issues¶

Ensure XDG_RUNTIME_DIR is writable:

ls -la /run/user/1000
# Should be owned by jovian:jovian with mode 700

No GPU Acceleration¶

Verify NVIDIA runtime: --gpus all
GPU is optional - pod works without it