Manifest V3, Explained

Manifest V3 is the new platform for Chrome extensions. It was announced in 2019, became the default for new extensions in 2021, and fully replaced Manifest V2 in 2024 across Chrome and Edge. By 2026, MV3 is just "how extensions work" on Chromium browsers. It is also the most controversial change in the history of browser extensions, particularly for ad blockers and privacy tools.

This article explains what changed, why it matters, who won and who lost, and what it means for the tools you use every day. No jargon, no axe to grind. For the bigger picture, see our best ad blocker 2026 round-up.

A short history

For most of the last decade, browser extensions on Chrome ran on Manifest V2, an API specification that gave extensions a lot of power. The most important power, for ad blockers, was the webRequest API. With webRequest, an extension could see every network request the browser made and, in real time, decide to allow it, block it, redirect it, or modify it.

That power was the reason classic uBlock Origin could do things no other blocker could: dynamic filtering, scriptlet injection, request modification based on context. It was also the reason a buggy or malicious extension could slow the browser to a crawl or, in the worst case, snoop on every page the user visited.

Manifest V3 reorganises this. Instead of granting extensions blanket runtime power, it asks them to declare what they want to do up front and lets the browser enforce those rules in native code.

What actually changed

From webRequest to declarativeNetRequest

The headline change. Under MV3, blockers no longer evaluate every request inside extension JavaScript. They submit a ruleset, in JSON, describing which requests to block, redirect or modify. The browser loads those rules once and applies them in native code on every request.

The upside: it is fast, predictable and easy to audit. The browser knows exactly what an extension can and cannot do because the rules are declared.

The downside: ruleset size is limited (currently 330,000 dynamic rules plus static rules from packaged lists), and rules are pattern-based, not arbitrary code. Advanced manoeuvres like "block this request only if the page already loaded a specific other resource" become harder or impossible.

Service workers instead of background pages

Extensions used to have a persistent background page that could run code as long as the browser was open. MV3 replaces this with service workers, which spin up only when needed and shut down after a few seconds of inactivity. This is great for memory but requires extensions to be carefully written, because anything that assumed long-lived state breaks.

Stricter content security policy

Inline scripts in extension pages are banned. Eval is banned. Remote code execution, even from your own server, is banned. Extensions must ship all the JavaScript they will ever run inside the package itself.

Host permissions are more granular

Users can grant an extension access to all sites, the current site, or specific sites, with clearer prompts. This is a privacy win.

Why ad blockers struggled

Most legacy blockers were built on the assumption that they had webRequest. Migrating to MV3 was not a tweak; it was sometimes a rewrite. Several casualties:

• Some blockers tried to ship literal ports of their MV2 logic and shipped with reduced filter coverage.
• Some abandoned features that worked beautifully under MV2 but had no MV3 equivalent.
• A few small blockers vanished entirely because the rewrite cost was too high for the team's resources.

The two open-source giants, uBlock Origin and AdGuard, navigated the transition differently. uBlock Origin's lead developer shipped uBlock Origin Lite as a deliberately scoped MV3 product and was vocal about what was lost. AdGuard, with more resources, ported aggressively and leaned on its desktop apps for the features the extension could no longer provide.

NovaBlock took a third path: build for MV3 from day one with no legacy to drag along. That choice is the reason the engine is small, the codebase is tidy, and the YouTube module can iterate independently of the filter lists. Background on the architecture in our how ad blockers work guide.

What MV3 means for you, the user

For most users, MV3 is invisible. You install an extension, it does its job, you do not think about the manifest version. The two practical changes you might notice:

1. Better defaults. Modern MV3 blockers spin up faster, use less memory and rarely have "wake up" pauses.
2. Slightly less power for hobbyists. If you used to write your own dynamic filter rules in classic uBlock Origin, you have probably already noticed those features are gone in uBlock Origin Lite.

For NovaBlock users, the change is invisible. The extension was designed for MV3, so there is nothing to "switch" or "upgrade". Filter lists are converted to declarativeNetRequest rules at install time and refreshed on a schedule. YouTube logic that cannot be expressed as a static rule lives in narrowly-scoped content scripts that run only on YouTube pages. Cookie banner suppression uses tightly targeted cosmetic rules and per-domain logic.

Performance, measured

Across our 25-site benchmark, well-built MV3 blockers actually outperform their MV2 predecessors:

The native-code rule evaluation is the biggest win. It is the reason your laptop fan does not spin up the way it sometimes did with older blockers.

Privacy implications

MV3's design genuinely improves the privacy of the extension platform. An MV3 extension has a narrower attack surface, cannot ship runtime code from a remote server, and has its rules visible and auditable.

The privacy of the blocker itself depends on the team, not the manifest. A determined, malicious MV3 extension can still misuse the host permissions a user granted it. The defence is the same as ever: install only extensions from teams you can vet, with clear privacy policies. Our chrome extension security article goes deep on how to evaluate.

Did Google have ulterior motives?

The cynical reading is that Google, an advertising company, benefits from constraints on ad blockers. The charitable reading is that Google, a browser vendor with billions of users, has legitimate reasons to want a safer extension platform. Both can be true at once.

What is undeniable is that the MV3 specification, as shipped, makes some advanced ad-blocking techniques impossible. It is equally undeniable that consumer-grade ad blocking, the kind 95 percent of users want, remains very effective. NovaBlock is proof of that: a polished consumer blocker, born on MV3, that misses none of the features a normal user actually uses.

Pros and cons of MV3 (for users)

Pros

• Faster, lower-memory extensions.
• Cleaner permission model, more user control.
• Smaller attack surface for malicious extensions.
• Easier to audit what an extension can do.

Cons

• Power-user filtering features lost in some products.
• Some legacy extensions abandoned rather than ported.
• A short transitional period of reduced coverage in late 2024.

What to do today

• Make sure your blocker is updated and explicitly MV3-compatible. If you are still running a 2023-era blocker that is no longer maintained, replace it.
• If you have not changed blockers in five years, consider switching. The MV3-native generation is meaningfully better.
• If you use Chrome and you want a blocker that was built for the modern platform from scratch, install NovaBlock.
• If you are a Firefox user, you have more flexibility. Firefox still supports webRequest, so classic uBlock Origin remains an option.

Conclusion

Manifest V3 is no longer a debate; it is the platform. The good news is that the doomy 2023 predictions did not come true. Effective ad and tracker blocking is alive and well under MV3, and in most cases, it is better. The blockers that survived the transition, and the ones built for MV3 from the start, are faster, lighter and easier to trust than the generation they replaced. NovaBlock is one of them, and the architecture story is told in more detail in our features and privacy pages.