Scrapybara

Scrapybara has shut down their virtual desktop and browser service as of October 15, 2025. If you were using Scrapybara for browser automation, Kernel is here to help you migrate seamlessly.

Key Concepts

Feature	Scrapybara	Kernel
Start Browser	client.start_browser()	kernel.browsers.create()
Standby Mode	instance.pause() / instance.resume()	Automatic standby mode
CDP URL	instance.get_cdp_url().cdp_url	Returns cdp_ws_url in create response
Live View	instance.get_stream_url().stream_url	Returns browser_live_view_url in create response
Stealth Mode	❌ Not available	Create browser with stealth: true
Replays	❌ Not available	kernel.browsers.replays.start() and kernel.browsers.replays.stop()
Save Auth	instance.browser.save_auth(name="default")	Create Profile. Then create browser with kernel.browsers.create(profile={"name": "profile1", "save_changes": true})
Click	instance.computer(action="click_mouse", button="left")	kernel.browsers.computer.click_mouse(id=session_id, x=100, y=200)
Drag	instance.computer(action="drag_mouse", path=[[100, 200], [300, 400]])	kernel.browsers.computer.drag_mouse(id=session_id, path=[[100, 200], [150, 220], [200, 260]])
Screenshot	instance.computer(action="take_screenshot").base64_image	kernel.browsers.computer.capture_screenshot(id=session_id)

How to migrate

Basic Browser Creation

Scrapybara

from scrapybara import Scrapybara
from playwright.async_api import async_playwright

client = Scrapybara(api_key="your_api_key")
instance = client.start_browser(timeout_hours=1)
cdp_url = instance.get_cdp_url().cdp_url

async with async_playwright() as p:
    browser = await p.chromium.connect_over_cdp(cdp_url)
    page = browser.contexts[0].pages[0]
    await page.goto("https://example.com")
    await browser.close()

instance.stop()

Kernel

from kernel import Kernel
from playwright.async_api import async_playwright

kernel = Kernel()
kernel_browser = kernel.browsers.create(timeout_seconds=3600)
cdp_url = kernel_browser.cdp_ws_url

async with async_playwright() as p:
    browser = await p.chromium.connect_over_cdp(cdp_url)
    page = browser.contexts[0].pages[0]
    await page.goto("https://example.com")
    await browser.close()

await kernel.browsers.delete_by_id(kernel_browser.session_id)

Save & Reuse Authentication

Scrapybara

# First session - save auth
instance = client.start_browser()
# ... login to website via Playwright ...
auth_state = instance.browser.save_auth(name="my-login")
instance.stop()

# Second session - load auth
instance2 = client.start_browser()
instance2.browser.authenticate(auth_state_id=auth_state.auth_state_id)
# ... browser now has saved cookies ...

Kernel

# First session - save auth
profile = await kernel.profiles.create(name="my-login")
kernel_browser = await kernel.browsers.create(
    profile={"name": "my-login", "save_changes": True}
)
# ... login to website via Playwright ...
await kernel.browsers.delete_by_id(kernel_browser.session_id)

# Second session - load auth
kernel_browser2 = await kernel.browsers.create(
    profile={"name": "my-login"}
)
# ... browser now has saved cookies ...

File Download

Scrapybara

instance = client.start_browser()
# ... trigger download in browser ...
# Then use file operations
downloaded_file = instance.file(command="read", path="/download/path")

Kernel

kernel_browser = await kernel.browsers.create()
# ... trigger download in browser via Playwright ...
# Then read from filesystem
file_response = await kernel.browsers.fs.read_file(
    kernel_browser.session_id,
    path="/tmp/downloads/file.pdf"
)
await file_response.write_to_file("local-file.pdf")

Long-Running Sessions

Scrapybara

# Pause/resume for long-running sessions
instance = client.start_browser(timeout_hours=24)
# ... do some work ...
instance.pause()  # Pause to save costs
# ... later ...
instance.resume()  # Resume work

Kernel

# Automatic standby mode + persistence
kernel_browser = await kernel.browsers.create(
    persistence={"id": "my-long-session"}
)
# ... do some work ...
# Browser enters standby mode between connections

# Later - reuse the exact browser instance with full state
kernel_browser2 = await kernel.browsers.create(
    persistence={"id": "my-long-session"}
)

Computer Controls

Both Scrapybara and Kernel provide Computer Controls APIs that allow you to programmatically control the browser environment at the system level - including mouse movements, clicks, keyboard input, and screenshots.

Scrapybara

instance = client.start_browser()

# Click at specific coordinates
instance.computer(action="click_mouse", button="right", coordinates=[300, 400])

# Drag from one position to another
instance.computer(action="drag_mouse", path=[[100, 200], [300, 400]])

# Type text
instance.computer(action="type_text", text="Hello World")

# Take a screenshot
screenshot = instance.computer(action="take_screenshot").base64_image

Kernel

kernel_browser = await kernel.browsers.create()

# Click at specific coordinates
kernel.browsers.computer.click_mouse(
    id=kernel_browser.session_id,
    x=100,
    y=200
)

# Drag from one position to another
kernel.browsers.computer.drag_mouse(
    id=kernel_browser.session_id,
    path=[[100, 200], [150, 220], [200, 260]],
    button="left",
    delay=0,
    steps_per_segment=10,
    step_delay_ms=50,
    hold_keys=["Shift"]
)

# Type text with optional delay
kernel.browsers.computer.type_text(
    id=kernel_browser.session_id,
    text="Hello World",
    delay=100
)

# Take a full screenshot
with open('screenshot.png', 'wb') as f:
    image_data = kernel.browsers.computer.capture_screenshot(id=kernel_browser.session_id)
    f.write(image_data.read())

For a complete reference of all available Computer Controls methods in Kernel, see the Computer Controls documentation.

Full API Comparison

Feature	Scrapybara	Kernel
Create Browser	client.start_browser()	kernel.browsers.create()
Get CDP URL	instance.get_cdp_url().cdp_url	Returns cdp_ws_url in create response
Get Live View	instance.get_stream_url().stream_url	Returns browser_live_view_url in create response
Delete Browser	instance.stop()	kernel.browsers.delete_by_id(session_id)
List Browsers	kernel.get_instances()	kernel.browsers.list()
Save Auth State	instance.browser.save_auth(name="default")	Create Profile. Then create browser with kernel.browsers.create(profile={"name": "profile1", "save_changes": True})
Load Auth State	instance.browser.authenticate(auth_state_id="xyz")	kernel.browsers.create(profile={"name": "profile1"})
Pause/Resume	instance.pause() / instance.resume()	Automatic standby mode
Screenshot	instance.screenshot()	Use Playwright or Puppeteer
Timeout Config	timeout_hours parameter	timeout_seconds parameter
Stealth Mode	❌ Not available	Create browser with stealth: true
Headless Mode	❌ Not available	Create browser with headless: true
Session Persistence	Auth state only	Full browser state via creating browser with persistence={"id": "persist-id"}
Video Replays	❌ Not available	kernel.browsers.replays.start() and kernel.browsers.replays.stop()
File Upload	instance.upload()	kernel.browsers.fs.upload() or Playwright
File Download	Via browser, then instance.file()	kernel.browsers.fs.read_file()
Process Control	instance.bash()	kernel.browsers.process.*
Proxy Support	❌ Not available	Create Proxy. Then create browser with kernel.browsers.create(proxy_id=proxy.id)
Click Mouse	instance.computer(action="click_mouse", button="left")	kernel.browsers.computer.click_mouse(id=session_id, x=100, y=200)
Move Mouse	instance.computer(action="move_mouse", coordinates=[100, 200])	kernel.browsers.computer.move_mouse(id=session_id, x=100, y=200)
Drag Mouse	instance.computer(action="drag_mouse", path=[[100, 200], [300, 400]])	kernel.browsers.computer.drag_mouse(id=session_id, path=[[100, 200], [150, 220], [200, 260]])
Scroll	instance.computer(action="scroll", coordinates=[100, 100], delta_x=0, delta_y=200)	kernel.browsers.computer.scroll(id=session_id, delta_x=0, delta_y=100)
Type Text	instance.computer(action="type_text", text="Hello")	kernel.browsers.computer.type_text(id=session_id, text="Hello")
Press Key	instance.computer(action="press_key", keys=["ctrl", "c"])	kernel.browsers.computer.press_key(id=session_id, keys=["Ctrl+t"])
Take Screenshot	instance.computer(action="take_screenshot").base64_image	kernel.browsers.computer.capture_screenshot(id=session_id)
Get Cursor Position	instance.computer(action="get_cursor_position").output	Use move_mouse with tracking

---

Need Help?

• Contact Us on Discord
• Sign-Up here
• Check out our open source repos here