SingLink Shield v1.2:Keeping Every UserStable and Secure

SingLinkVPN’s security upgrade went live on November 1. We still keep no logs, and an independently trained safety model stops malicious behavior and hacker attacks in real time.

  • 99.99%of users unaffected
  • 0.01%malicious violators restricted
  • 1%misjudgement rate, at most
  • v1.2live since November 1

Key facts

What it is
SingLink Shield is SingLinkVPN’s security system. Version v1.2 has been in effect since November 1.
Safety model
Independently trained by SingLinkVPN together with a third-party security company, using leading security technology.
Your data
No logs are kept, and no SingLinkVPN user data is used to train the model. SingLink Shield runs in the app and needs no data to be retained or read.
Who is affected
99.99% of users are not affected; measures apply only to the 0.01% of malicious users.
Misjudgements
The misjudgement rate is as low as 1%. Misjudgements are lifted quickly and automatically, and you can appeal at appeal.singlinkvpn.com.
Three security models
Stopping malicious behavior, stopping hacker attacks and upholding the Terms of Service.
  • Strictly no logs
  • An independently trained safety model
  • Real-time decisions and action
  • Low misjudgement, open to appeal
  • Stopping malicious behavior
  • Stopping hacker attacks
  • Upholding the Terms of Service

Why we upgradedSingLink Shield

Security incidents have been frequent around the world recently, and software security can’t be ignored. As AI takes off, attack methods are becoming more varied.

To meet growing security threats and keep every connection stable and secure, SingLinkVPN has upgraded its security system and related protections. We take action against users who try to reverse-engineer, crack, attack or threaten the service, including stopping their use of the service and terminating their account access.

Technical whitepaper:how SingLink Shield works

A plain-language look at the architecture behind SingLink Shield: three lines of defence, one safety model, and a decision loop that includes automatic lifting and appeals.

Three security models

SingLink Shield is built on three security models, each aimed at a different risk.

  1. Stopping malicious behavior

    Targets reverse engineering, cracking, abuse and maliciously hogging bandwidth. Security measures are taken mainly against malicious users, such as suspending their use of the SingLinkVPN service.

  2. Stopping hacker attacks

    Deploys security protections against hackers and all kinds of malicious attacks.

  3. Upholding the Terms of Service

    Takes protective measures against behavior that violates the Terms of Service, such as suspending use of the SingLinkVPN service.

From detection to correction

  1. Identify

    The safety model identifies harmful behavior, such as reverse engineering, cracking, abuse or attacks.

  2. Decide in real time

    Rule violations are judged in real time.

  3. Take measures

    Appropriate measures are taken, such as suspending use of the service or terminating account access.

  4. Correct and appeal

    Misjudgements are lifted quickly and automatically, and you can appeal at appeal.singlinkvpn.com.

Privacy by design

  • Before and after the upgrade, SingLinkVPN keeps no logs of any kind.
  • SingLink Shield runs in the app: a built-in recognition system automatically identifies abnormal behavior and acts in real time, without retaining or reading any data.
  • The safety model is trained without any SingLinkVPN user data.
  • Training follows security-compliant methods provided by a third-party security company.

How the safetymodel is trained

The SingLink Shield safety model was independently trained by SingLinkVPN together with a third-party security company, using leading security technology.

  1. Trained with a third-party security company

    SingLinkVPN and a third-party security company trained the model together and independently.

  2. No user data

    Training keeps no logs and uses no SingLinkVPN user data at all.

  3. Security-compliant methods

    Training follows security-compliant methods provided by the third-party security company.

  4. Simulation and long-term training

    After many simulations and long-term training, the misjudgement rate is as low as 1%.

Continuous optimisation

Current version

v1.2 · in effect since November 1. The safety model is continuously optimised.

How SingLinkShield works

SingLink Shield protects every user without compromising your privacy.

  1. Strictly no logs

    As always, SingLinkVPN keeps no logs of any kind, protecting your privacy and security.

  2. An independently trained safety model

    An independently trained safety model identifies harmful behavior.

  3. Real-time decisions and action

    Rule violations are judged in real time, and appropriate measures are taken.

  4. Low misjudgement, open to appeal

    The misjudgement rate is as low as 1%. Misjudgements are lifted automatically, and you can also appeal.

Real-worldscenarios

What SingLink Shield does in everyday situations, and what, if anything, you need to do.

Same connection.Three cases.

SingLink Shield only targets malicious behavior. Here is what different users can expect under SingLink Shield v1.2.

  1. Regular users

    Affected?
    Not affected: 99.99% of users are in this group today.
    What SingLink Shield does
    You connect as usual, and SingLinkVPN still keeps no logs.
    How it is resolved
    Nothing to do.
  2. Misjudged users

    Affected?
    The misjudgement rate is as low as 1%, and the restriction is lifted quickly and automatically.
    What SingLink Shield does
    It automatically lifts the restriction caused by the misjudgement.
    How it is resolved
    If you disagree, submit an appeal on the appeal site or contact online support.
  3. Malicious violators

    Affected?
    Restricted. They make up only 0.01% of users.
    What SingLink Shield does
    It takes measures such as suspending the service or terminating account access.
    How it is resolved
    Behavior that violates the Terms of Service is met with protective measures.

Almost no oneis affected

At present, 99.99% of users are not affected. Restrictions apply only to the 0.01% who are malicious users.

SingLink Shield has effectively stopped people who try to undermine other users’ stability and security. Our goal hasn’t changed: a more stable, more secure experience for every user.

Who counts as a violator?

Generally, anyone who tries to affect SingLinkVPN through malicious means such as reverse engineering, cracking, attacks or threats.

Think you were misjudged?Appeal

Our safety model has been through independent simulation training, and its misjudgement rate is as low as 1%. Even if you are misjudged, the restriction is lifted quickly and automatically.

If you disagree with a decision, submit an appeal on the SingLinkVPN appeal site. You can also contact online support.

Heavy daily data use? Choose Pro

If you need large amounts of VPN data every day over a long period, for example downloading thousands of GB a day, please subscribe to the Pro plan. Plus is designed for everyday professional work, not large-volume downloads.

See the Pro plan

Frequently askedquestions

01What is SingLink Shield?

SingLink Shield is SingLinkVPN’s security system. An independently trained safety model identifies harmful behavior in real time, and the system is built on three security models: stopping malicious behavior, stopping hacker attacks and upholding the Terms of Service. Version v1.2 has been in effect since November 1.

02Does SingLink Shield record or read my data?

No. SingLinkVPN still keeps no logs of any kind. SingLink Shield runs in the app: its built-in recognition system automatically identifies abnormal behavior and takes action in real time, and the whole process requires no data to be retained or read.

03How is the safety model trained?

It was independently trained by SingLinkVPN together with a third-party security company, using leading security technology and security-compliant methods provided by that company. After many simulations and long-term training, the misjudgement rate is as low as 1%.

04Is my data used to train the model?

No. Training keeps no logs and does not use any SingLinkVPN user data.

05Who gets restricted?

Only the 0.01% of malicious users who try to affect SingLinkVPN through reverse engineering, cracking, attacks or threats, or who violate the Terms of Service. Measures include suspending use of the service and terminating account access.

06Will this affect me as a regular user?

No. At present, 99.99% of users are not affected.

07What if I am misjudged, and how long does it take?

The misjudgement rate is as low as 1%, and misjudgements are lifted quickly and automatically.

08How do I appeal?

Submit an appeal at appeal.singlinkvpn.com. You can also contact online support.

09Why isn’t Plus suitable for heavy downloads?

Plus is designed for everyday professional work. If you download thousands of GB a day over a long period, please choose the Pro plan.

10What is v1.2, and will SingLink Shield be updated?

v1.2 is the current version of SingLink Shield, in effect since November 1. The safety model is continuously optimised.

Last updated: