Valve's VACNet Labeling Portal Explained: How Community Reviewers Now Train CS2's Anti-Cheat

In August 2026, Valve quietly began testing a new anti-cheat tool for Counter-Strike 2 - the VACNet Labeling Portal. The system resembles the Overwatch feature familiar from CS:GO: selected users once again review gameplay clips and evaluate suspicious behavior. However, the purpose of the new portal is noticeably different.

The main change is that a reviewer’s decision is not supposed to directly result in a player being banned. Instead, human evaluations are used as labeled data that may help train future VACNet models.

What Is the CS2 VACNet Labeling Portal?, How Does the New VACNet Portal Work?, How Is the New Portal Different From the Old Overwatch?, Why Does VACNet Need Human Reviews?, Who Can Get Access?, Why Does Valve Ask Users Not to Share Clips?, Will VACNet Solve CS2's Cheating Problem?

What Is the CS2 VACNet Labeling Portal?

The first signs of the system appeared after a small CS2 update released during the night of August 10-11, 2026. The patch was around 42 MB in size. Game files contained elements related to the VACNet Labeling Portal, including a new icon and a user access check.

For players who receive access, a small “V” icon appears in the CS2 main menu. Instead of opening a regular in-game menu, it takes the user to a separate page on the official Counter-Strike website.

The portal itself is hosted on Valve’s Counter-Strike.net domain. Its public page explicitly describes it as the Counter-Strike 2 VACNet labeling portal, while access is limited to users invited to participate.

This means the feature is not simply something discovered by dataminers, but an active Valve service with restricted access.

How Does the New VACNet Portal Work?

Once access is granted, the user is shown Counter-Strike 2 match clips and asked to evaluate the actions of a selected player.

According to early participants in the testing, the available categories include:

• aim assistance

• gaining information about opponents through walls

• automated movement such as bunnyhopping

• fully automated accounts, such as farming bots.

The reviewer watches a gameplay clip with sound and then applies the appropriate labels. The available options also allow users to indicate that it is impossible to confidently determine whether a violation occurred.

Valve specifically emphasizes that choosing an uncertain option is perfectly acceptable. Reviewers are not expected to accuse or clear a player at any cost - it is more important to provide a confident assessment only when the behavior is actually clear.

This is an important difference from a regular in-game report. The reviewer is effectively acting as a data labeler.

How Is the New Portal Different From the Old Overwatch?

In CS, Overwatch was primarily a community-based system for reviewing suspected violations. Experienced players watched demos of suspected cheaters and submitted verdicts. The combined decisions could then be used as part of the punishment process.

The VACNet Labeling Portal is built around a different idea.

Valve tells invited users that their decisions do not directly affect the player being reviewed. Instead, the labels they create may be used to train future VACNet models.

That is why calling the new system simply “the return of Overwatch” would not be entirely accurate. Visually and conceptually, the similarities are clear - a human is once again analyzing suspicious gameplay. But the main purpose now is not to pass judgment on a specific account, but to create high-quality data for machine learning.

Why Does VACNet Need Human Reviews?

The idea itself is not new. Back in 2018, Valve engineer John McDonald spoke in detail at GDC about using deep learning to detect cheating behavior in Counter-Strike.

Valve trained the system using gameplay data and human evaluations, then regularly updated its models. In earlier versions of VACNet, machine learning made it possible to analyze behavior directly in matches without relying only on client-side detection of cheat software.

The new portal effectively brings human labeling back into that cycle.

In simplified form, the process looks like this: the system receives a gameplay clip, a person identifies the type of behavior they see, the clip receives a label, and the accumulated set of examples may later be used to further train VACNet.

At the same time, Valve has not revealed the current VACNet architecture, dataset size, or decision-making process in CS2. Because of that, it is not possible to say that every submitted label is automatically used for training or directly influences a specific version of the anti-cheat system.

Who Can Get Access?

This remains the biggest unanswered question.

Valve has not published any official participation requirements. There is no confirmed minimum Premier Rating, required number of hours, number of wins, or mandatory Trust Factor level. Valve has also not provided a public application form that guarantees access to the program.

Gabe Follower reported on one player who received access to the portal. According to him, the user had around 16,000 Premier Rating and had previously participated in Overwatch in CS. However, this is only one specific example and not a confirmed set of Valve requirements.

For that reason, advice such as “reach a certain rating and you will unlock VACNet” currently has no reliable confirmation.

Why Does Valve Ask Users Not to Share Clips?

Portal participants are also asked not to stream, record, or distribute the clips they review.

The reason is fairly straightforward: publicly sharing a large number of examples could show cheat developers exactly what type of behavior is being flagged for review. Valve directly links this restriction to avoiding giving bad actors more information about how to better hide cheating activity.

This is especially important for the new system. If the purpose of the portal is to improve detection models, keeping the details of the review process private reduces the amount of feedback that cheat developers could potentially use to create better bypasses.

Best CS2 Cases for Knives in 2026: Odds & Cost-Per-Knife Compared

Will VACNet Solve CS2's Cheating Problem?

The introduction of the Labeling Portal looks like an important step, but it is still too early to treat it as an immediate solution to cheating in CS2.

The portal does not mean that every suspected cheater being reviewed will instantly receive a ban. In fact, Valve explicitly tells reviewers that their decisions do not directly affect the specific player they are watching.

The potential benefit is more long-term. The more high-quality examples of legitimate and suspicious behavior the system receives, the better future models may become at automatically distinguishing between them.

That is what makes the VACNet Labeling Portal more interesting than a simple return of Overwatch. Valve appears to be using community experience not only to review individual cases, but also to help scale automated detection of suspicious behavior across the entire game.

While Valve continues working on match integrity and training VACNet, you can take a short break from matchmaking - for example, by opening a case on Farmskins, without having to depend on the randomness of your teammates in yet another lobby.