apify/crawlee · error · NavigationSkippedError

The `window` property is not available - `skipNavigation` wa

Error message

The `window` property is not available - `skipNavigation` was used

What it means

LinkedomCrawler mirrors JSDOMCrawler's skipNavigation design: with navigation skipped, no parse5/linkedom window is built, and the `window` property getter throws NavigationSkippedError to signal that DOM access is not possible in this mode.

Source

Thrown at packages/linkedom-crawler/src/internals/linkedom-crawler.ts:248

                crawlingContext.body.toString(),
                isXml ? 'text/xml' : 'text/html',
            );

            return {
                window: document.defaultView,
                get body() {
                    return document.documentElement.outerHTML;
                },
                get document() {
                    // See comment about typing in LinkeDOMCrawlingContext definition
                    return document as unknown as Document;
                },
            };
        } catch (err) {
            if (err instanceof NavigationSkippedError) {
                return {
                    get window(): Window {
                        throw new NavigationSkippedError(
                            'The `window` property is not available - `skipNavigation` was used',
                            { cause: err },
                        );
                    },
                    get body(): string {
                        throw new NavigationSkippedError(
                            'The `body` property is not available - `skipNavigation` was used',
                            { cause: err },
                        );
                    },
                    get document(): Document {
                        throw new NavigationSkippedError(
                            'The `document` property is not available - `skipNavigation` was used',
                            { cause: err },
                        );
                    },
                };
            }

View on GitHub (pinned to dbe57fb09c)

Solutions

Remove `skipNavigation: true` if DOM/window access is needed.
Use `context.body` (raw string) and parse with cheerio or regex instead.
Branch the handler based on navigation-skip state before touching `window`.
Use extractLinks only on requests where the window was actually created.

Example fix

// before (skipNavigation: true)
const links = [...context.window.document.querySelectorAll('a')];

// after
const $ = (await import('cheerio')).load(context.body);
const links = $('a[href]').toArray();

Defensive patterns

Strategy: try-catch

Validate before calling

if (request.skipNavigation === true) {
    // context.window will throw; use cheerio on context.body instead
}

Type guard

function isNavigationSkippedError(err: unknown): err is Error & { name: 'NavigationSkippedError' } {
    return err instanceof Error && err.name === 'NavigationSkippedError';
}

Try / catch

try {
    const { window } = context;
    parse(window);
} catch (err) {
    if ((err as Error).message.includes('`skipNavigation` was used')) {
        parseCheerio(context.body);
    } else {
        throw err;
    }
}

Prevention

Guard all window access behind a skipNavigation check.
Use separate handlers for skipNavigation vs parsing crawls.
Store skipNavigation flag in request.userData for easy branching.
Add tests that exercise handlers in both navigation modes.

When it happens

Trigger: Using LinkedomCrawler with `skipNavigation: true` and then reading `context.window` in the requestHandler.

Common situations: Performance-tuned linkedom crawls with skipNavigation but handlers that still call window-based helpers (enqueueLinks, extractLinks); handlers shared between linkedom and normal crawlers.

Related errors

AI-assisted analysis of apify/crawlee@dbe57fb09c (2026-08-30). Data as JSON: /api/errors/9b4691d0345ee6f9. Report an issue: GitHub.