Remove ad-like DOM elements in Chrome

01 · Role in the system

What Jev does here

Jev answers one Noul question per compact candidate description in a batch. Candidate discovery, page-size guards, thresholding, animation, removal, and mutation observation are ordinary extension code. The project is explicitly a demonstration rather than a replacement for a security or privacy-focused ad blocker.

Heuristics nominate, one batched judgment decides

The extension's pipeline keeps the model's workload minimal: deterministic heuristics find ad-shaped DOM candidates, each candidate becomes one compact description, and a single batched request answers one yes-or-no question per candidate — is this a paid advertisement? Page-size guards, thresholding, removal, and mutation observation are all ordinary extension code.

The one-judgment-per-candidate shape has a nice property worth noticing: the question is identical across candidates, so a misjudgment is visible as a threshold problem, not a prompt problem. Tuning the extension means moving a bar in code, not editing instructions and hoping.

A demonstration that declines to be a product

The record is explicit that this is a demonstration, not a replacement for a security or privacy-focused blocker: it can miss ads, it can remove non-ad content, and it does not touch tracking, malware, or video advertising. Because only heuristic candidates reach the model, ads the heuristics never surface are invisible to the whole pipeline — a boundary the limitations state clearly.

Within that scope, the record is a clean example of judgment-augmented DOM manipulation: candidate generation stays deterministic and auditable, the model contributes only a classification that code enforces, and the failure mode is a wrongly removed node that a user can see, not a silent data flow.

02 · Control boundary

Where Jev sits

Heuristic DOM candidates to batched advertisement probabilities; browser code thresholds and removes matching nodes.

Code owns the loop, permissions, thresholds, validation, and side effects. Jev owns only the bounded judgments described above.

03 · Known limits

What this evidence does not prove

• It can miss ads or remove non-ad content and does not block tracking, malware, or video advertising.
• Only candidates found by deterministic heuristics are sent for judgment, so hidden or unlabeled ads can be missed.

04 · Attribution

Public sources

This is a Community record: the project was published by a third-party community author.

• realZachi ↗Community · github · public · checked 2026-09-18