How to Take a Screenshot in Vibium

To take a screenshot in Vibium, call screenshot() to get PNG bytes and write them to a file — for example png = vibe.screenshot() then open("out.png", "wb").write(png). It works on the whole page or on a single element: call screenshot() on the browser/page object for a full-page capture, or on an element returned by find() to capture just that region. Vibium returns the image as raw PNG bytes rather than writing a file for you, which makes it easy to stream the screenshot elsewhere — into a visual-diff tool, an S3 upload, or a vision model that an AI agent uses to verify the page. Because Vibium auto-waits for actionability, the element is ready before you capture it, so screenshots don't come out half-rendered or flaky. The complete, runnable patterns for both the full-page and single-element cases are below.

Take a full-page screenshot

from vibium import browser_sync as browser
 
vibe = browser.launch()
vibe.go("https://example.com")
 
png = vibe.screenshot()
with open("screenshot.png", "wb") as f:
    f.write(png)
 
vibe.quit()

screenshot() returns the image as PNG bytes. You write those bytes to a .png file yourself — Vibium doesn't hide the data behind a path, which makes it easy to stream the image somewhere else (an AI agent, an S3 upload, a visual-diff tool).

Screenshot a single element

Find the element first, then call screenshot() on it to capture only that region:

from vibium import browser_sync as browser
 
vibe = browser.launch()
vibe.go("https://example.com")
 
header = vibe.find("h1")
png = header.screenshot()
with open("header.png", "wb") as f:
    f.write(png)
 
vibe.quit()

Common uses

• Visual regression — capture a baseline image and diff it on every run.
• Debugging failures — screenshot the page when an assertion fails.
• Feeding AI agents — hand the PNG bytes to a vision model for verification.

Next steps

• Automate a login flow
• What is Vibium?