nokap

nokap captures screenshots and PDFs from web pages (or local HTML) using headless Chrome via the Chrome DevTools Protocol. It doesn’t need Selenium or Playwright. Its only runtime dependencies are websockets and click.

Install and Go

Get nokap from PyPI and start capturing screenshots immediately:

pip install nokap

nokap finds the Chrome or Chromium already on your system. If there isn’t one, or you’re working inside a sandbox where full Chrome can’t start, run nokap install-chrome to download Chrome’s standalone headless build (see Installation).

Then in Python:

import nokap

# Screenshot a URL
nokap.webshot("https://example.com", "example.png")

# From HTML string (useful for table libraries)
html = "<h1>Hello, world!</h1>"
nokap.from_html(html, "hello.png")

Or from the command line:

nokap webshot https://example.com screenshot.png
nokap from-html report.html report.png --selector "table" --zoom 2

Why nokap?

Zero heavyweight deps	No Selenium, no Playwright: just `websockets`, `click`, and Chrome
Screenshots	PNG, JPEG, WebP via CDP `Page.captureScreenshot`
PDF generation	Configurable page size, margins, orientation
CSS selectors	Capture specific elements or unions of multiple selectors
Retina quality	Zoom/scale for 2× or 3× resolution images
CLI included	`nokap webshot` and `nokap from-html` commands
Jupyter safe	Works in notebooks without event loop conflicts
Sandbox friendly	Runs in macOS agent sandboxes with `chrome-headless-shell`

Configuration

Environment Variable	Purpose
`CHROME_PATH`	Override Chrome binary location
`NOKAP_CHROME_ARGS`	Extra Chrome command-line flags for every launch
`NOKAP_CACHE_DIR`	Where `nokap install-chrome` stores browsers

Requirements

Python ≥ 3.10
Chrome or Chromium installed on the system, or chrome-headless-shell installed with nokap install-chrome

License

MIT

Links

AI / Agents

Developers

Community