Why doesn't ||notrack.ai^$document block notrack.ai?

[Stevoisiak]

I'm trying to use $document to block the notrack.ai website from loading. However, for some reason the filter ||notrack.ai^$document isn't blocking it.

The uBO logger doesn't list any doc requests from notrack.ai.

By contrast, the filter does block the site in Adguard.

Is this expected behavior? Do some sites not send $doc requests? If yes, is there a standard way of blocking those sites?

Edit: Troubleshooting info. Am using Microsoft Edge on Windows 11.

uBlock Origin: 1.75.0
Chromium: 156
filterset (summary):
 network: 1
 cosmetic: 0
 scriptlet: 0
 html: 0
listset (total-discarded, last-updated):
 removed:
  ublock-filters: null
  ublock-badware: null
  ublock-privacy: null
  ublock-unbreak: null
  ublock-quick-fixes: null
  easylist: null
  easyprivacy: null
  urlhaus-1: null
  plowe-0: null
 default:
  user-filters: 1-0, never
filterset (user): [array of 1 redacted]
userSettings:
 advancedUserEnabled: true
 userFiltersTrusted: true
hiddenSettings:
 trustedListPrefixes: ublock- https://raw.githubusercontent.com/Stevoisiak/
supportStats:
 allReadyAfter: 618 ms (selfie)
 maxAssetCacheWait: 259 ms
 cacheBackend: indexedDB

Video:

2026-10-27.20-20-52-00.00.00.000-00.00.45.462.mp4

[stephenhawk8054]

It works for me:

Details

You need to describe how to reproduce the issue.

[gothic-bum]

Perform a page reload that ignores cached content:

Ctrl + Shift + R

OR

Ctrl+F5

[Stevoisiak]

Strangely, the block initially worked when I did a non-cached reload, but then when I reloaded a second time it no longer worked.

[gothic-bum]

Try testing using a new Edge profile with only uBO and its default settings + the custom filter ||notrack.ai^$document

[freezer2022]

I can't reproduce either.

From what can be found, similar cases occurred only on SPA websites:

• Strict blocking does not work on Youtube (History.pushState) #3119

(but it doesn't seem your case to be SPA, as you clearly mentioned "I reloaded").

[Stevoisiak]

I did a bit more testing. The block does seem to work when doing a hard refresh. However, it does not work when entering the page URL directly into the address bar, even after doing a refresh.

2026-10-29.11-04-34.mp4

[gorhill]

The uBO logger doesn't list any doc requests from notrack.ai

See if the main document is shown in the Network pane in dev tools:

[Stevoisiak]

Yes, the document is shown in the devtools log. I can only reproduce this issue on 1 of my two PCs.

[gorhill]

By the look of it, the main document was fetched through the service worker on your side, that's the difference I see from my side.

[freezer2022]

Maybe this solution can work:

#2833 (comment): Try: ||notrack.ai^$csp=worker-src 'none'
and unregister service worker from about:serviceworkers.

As a source of information, there exist several threads about various uBO filters not working due to the content being served by the Service Worker:

• HTML Filter doesn't work on service worker requests #2833
• Workaround wrong webrequest-initiators from serviceworkers' fetch (chrome-only) #2207
• Does not block large media fetched over Fetch API #610
• Fetch API in service worker fails on homepage even with domain whitelisted #173