ScrapeGraphAI/Scrapegraph-ai · error · ImportError

The dependencies for screenshot scraping are not installed.

Error message

The dependencies for screenshot scraping are not installed. Please install them using `pip install scrapegraphai[ocr]`.

What it means

Raised by take_screenshot when PIL (Pillow) is not importable, meaning the optional screenshot-scraping dependencies are missing. The message points to the 'ocr' extra. The original ImportError is chained.

Source

Thrown at scrapegraphai/utils/screenshot_scraping/screenshot_preparation.py:28

from ..logging import get_logger

logger = get_logger(__name__)


async def take_screenshot(url: str, save_path: str = None, quality: int = 100):
    """
    Takes a screenshot of a webpage at the specified URL and saves it if the save_path is specified.
    Parameters:
        url (str): The URL of the webpage to take a screenshot of.
        save_path (str): The path to save the screenshot to. Defaults to None.
        quality (int): The quality of the jpeg image, between 1 and 100. Defaults to 100.
    Returns:
        PIL.Image: The screenshot of the webpage as a PIL Image object.
    """
    try:
        from PIL import Image
    except ImportError as e:
        raise ImportError(
            "The dependencies for screenshot scraping are not installed. "
            "Please install them using `pip install scrapegraphai[ocr]`."
        ) from e

    async with async_playwright() as p:
        browser = await p.chromium.launch(headless=True)
        page = await browser.new_page()
        await page.goto(url)
        image_bytes = await page.screenshot(
            path=save_path, type="jpeg", full_page=True, quality=quality
        )
        await browser.close()
        return Image.open(BytesIO(image_bytes))


def select_area_with_opencv(image):
    """
    Allows you to manually select an image area using OpenCV.

View on GitHub (pinned to 532dfffbf6)

Solutions

1. pip install 'scrapegraphai[ocr]' (or uv sync --extra ocr)
2. Or install Pillow directly: pip install Pillow
3. Verify playwright browsers are installed (playwright install chromium) if the next step fails
4. Pin compatible Pillow version if there's a conflict

Example fix

# before
# take_screenshot(url) -> ImportError
# after
pip install 'scrapegraphai[ocr]'

try:
    import PIL  # noqa
    deps_ok = True
except ImportError:
    deps_ok = False

try:
    img = await take_screenshot(url)
except ImportError as e:
    raise RuntimeError("install extras: pip install 'scrapegraphai[ocr]'") from e

When it happens

Trigger: Calling take_screenshot in an environment where Pillow (or the playwright stack bundled with the ocr extra) is not installed.

Common situations: Base install of scrapegraphai without extras; CI images trimmed of optional deps; Pillow removed accidentally or broken by a version conflict.

Related errors

• module {__name__!r} has no attribute {name!r}
• Could not import the 'ddgs' package required for DuckDuckGo

AI-assisted analysis of ScrapeGraphAI/Scrapegraph-ai@532dfffbf6 (2026-08-28). Data as JSON: /api/errors/6a4e579624474e42. Report an issue: GitHub.