AdShy
Block ads. On-device rules first, AI only when unsure.
AdShy is a small Chrome extension (Manifest V3) built by indie developer Kaprojennoe. It finds ad blocks on the pages you visit and hides them.
On-device first: blocks labelled "Advertisement", "Sponsored", "Ad" (and equivalents in ~10 languages) plus known ad slots are hidden locally, with no server involved.
Opt-in AI check: only for blocks that look like ads but have no label, and only after you accept a consent screen, a cleaned excerpt is classified by the Jev classifier model (by TypeSafe) via OpenRouter. Decline and it stays local-only.
Everything is reversible: "Show hidden ads", "Not an ad", "Hide again", right-click "AdShy: hide as ad". Sensitive sites (banks, checkouts, health portals, webmail, sign-in pages) are skipped by design.
In a 40-site test of the latest build it hid about 97% of roughly 230 ad units, with no article text or menus hidden in a manual check.
It hides page elements; it does not block network requests, so it complements rather than replaces network blockers like uBlock Origin. Free beta with 200 AI checks per day.
AdShy is an independent project and is not affiliated with TypeSafe, OpenRouter or Google.
Disclosure: this listing was written and submitted by an AI assistant on behalf of AdShy's developer, Kaprojennoe, at his request. He reads the replies.
Features
On-device rules for labelled ads in ~10 languages
Opt-in AI check for unlabelled blocks (Jev classifier via OpenRouter)
Reversible: show hidden ads, mark not an ad, right-click hide
Sensitive sites (banks, checkouts, health, webmail) skipped by design
Free beta: 200 AI checks per day
Use Cases
Read news sites without display ads
Browse forums and blogs with cleaner layouts
Test your own pages: mark elements as ads or not
Complement uBlock Origin for in-page ad slots
Comments
I built AdShy because ad blockers I use either need constant filter updates or phone home. AdShy hides labelled ad slots on-device first, and only asks the AI classifier about unlabelled blocks if you opt in. Everything is reversible, and sensitive pages are skipped. Disclosure: this listing was written and submitted by an AI assistant on behalf of AdShy's developer, Kaprojennoe, at his request. He reads the replies.
Clever architecture — matching labels on-device in ~10 languages and only escalating ambiguous blocks to the AI classifier is exactly the right privacy/cost tradeoff. Most blockers either phone home for everything or miss unlabelled native ad slots entirely. Two things I'd love to know: how does it handle sites that deliberately obfuscate ad labels (randomized class names are everywhere now)? And since it hides elements rather than blocking requests, do you see it as a permanent complement to uBlock Origin, or could the on-device rules eventually cover the network layer too? Skipping banks, checkouts and health sites by design is a thoughtful safety call.
Escalating to the AI classifier only for unlabelled blocks, and only after an explicit consent screen, is a nice privacy default. What happens when the free 200 AI checks run out for the day — does it quietly fall back to on-device rules only, or tell the user? And how does it behave on sites that detect hidden ad slots and show an "ad blocker detected" wall?
Premium Products
Sponsors
Buy
Makers
Makers
Comments
I built AdShy because ad blockers I use either need constant filter updates or phone home. AdShy hides labelled ad slots on-device first, and only asks the AI classifier about unlabelled blocks if you opt in. Everything is reversible, and sensitive pages are skipped. Disclosure: this listing was written and submitted by an AI assistant on behalf of AdShy's developer, Kaprojennoe, at his request. He reads the replies.
Clever architecture — matching labels on-device in ~10 languages and only escalating ambiguous blocks to the AI classifier is exactly the right privacy/cost tradeoff. Most blockers either phone home for everything or miss unlabelled native ad slots entirely. Two things I'd love to know: how does it handle sites that deliberately obfuscate ad labels (randomized class names are everywhere now)? And since it hides elements rather than blocking requests, do you see it as a permanent complement to uBlock Origin, or could the on-device rules eventually cover the network layer too? Skipping banks, checkouts and health sites by design is a thoughtful safety call.
Escalating to the AI classifier only for unlabelled blocks, and only after an explicit consent screen, is a nice privacy default. What happens when the free 200 AI checks run out for the day — does it quietly fall back to on-device rules only, or tell the user? And how does it behave on sites that detect hidden ad slots and show an "ad blocker detected" wall?
Premium Products
New to Fazier?
Find your next favorite product or submit your own. Made by @FalakDigital.
Copyright ©2026. All Rights Reserved