AdGuard VPN Review (2026): Open Protocol, Unaudited No-Logs Claim

Quick answer

AdGuard VPN still has no published independent no-logs audit. The Android app passed a 2025 Mobile App Security Assessment, and AdGuard released reference code for its TrustTunnel protocol in January 2026. Both improve the evidence. Neither shows what production VPN servers retain.

As of 26 July 2026, the direct EUR price was EUR 10.79 monthly, EUR 3.59 per month billed yearly, or EUR 2.69 per month billed every two years, before applicable VAT. The current matrix score remains 2.88. That is a defensible middle result for an approachable, broad-platform product whose technical transparency has improved but whose central privacy promise remains self-reported.

Has AdGuard VPN been audited?

AdGuard VPN has not published an independent report testing its no-logs claim, production server configuration or data-retention controls. Its current privacy policy makes the claim but cites no outside server audit (privacy policy).

There is one real outside assessment, and its scope matters. In 2025 the Android app passed a Mobile App Security Assessment by Leviathan Security Group and earned Google Play's independent-security-review badge (AdGuard's account). That is evidence about the mobile application. AdGuard has not published the findings report, and the assessment does not establish server-side non-retention.

The direct answer is therefore: the Android app was assessed; the no-logs promise was not audited. Treating the badge as a no-logs audit would answer a larger question than the evidence supports.

TrustTunnel is now open source

The old description of AdGuard VPN as using a closed protocol is now wrong. On 21 January 2026 AdGuard published TrustTunnel's reference client and server and said the same technology runs in its VPN apps (announcement; source repository).

TrustTunnel carries traffic through TLS using HTTP/2 or HTTP/3, with TCP, UDP and ICMP support. Fact: the code is public. Provider report:AdGuard says the web-like transport helps on restricted networks. Unresolved:we found no published independent protocol audit or comparative censorship test proving how reliably that claim holds.

Open source makes the design inspectable. It does not automatically make a young custom protocol as battle-tested as WireGuard or OpenVPN, and it says nothing by itself about server logging. Readers in restrictive jurisdictions should also check local law; this review does not recommend unlawful evasion.

Who operates and controls AdGuard VPN

The privacy policy names ADGUARD SOFTWARE LIMITED, registered in Limassol, Cyprus, as data controller. The about page says that company develops AdGuard VPN, AdGuard Ad Blocker and AdGuard DNS. It also says three founders head the business, but does not name them or publish a shareholding chain (privacy policy; about page).

The precise conclusion is that the operator and controller are verified; ultimate beneficial ownership is not resolved by the current public first-party record. Absence of a displayed parent is not proof of independence. We corrected that label in our VPN ownership map.

What AdGuard VPN collects

AdGuard says its VPN servers do not record browsing activity or users' originating IP addresses. The account layer is not data-free. The policy says it processes account email, payment records through Paddle or PayPro, active connection-instance counts, and traffic volume per account. Traffic-volume records are kept for 90 days without location data to administer limits. Diagnostics and usage metrics are optional.

This is not evidence that AdGuard records browsing. It is evidence that "no logs" is too broad a phrase. The relevant claim is narrower: no activity logs on VPN servers. Until an auditor tests the production controls, that remains an interested company statement. Our methodology distinguishes stated policy from independently verified practice.

Current price and app versions

On 26 July 2026 the direct EUR purchase page showed EUR 10.79 for one month, EUR 3.59 per month billed annually and EUR 2.69 per month billed every two years. VAT may apply. The one- and two-year plans include a 30-day money-back guarantee. Paid accounts cover up to 10 devices. The free tier offers 3 GB each month and two devices.

Current stable releases listed by AdGuard were Windows 2.9.6 (7 July), macOS 2.9.2 (29 June), Android 2.16.1 (16 June), iOS 2.10.1 (16 June), browser extension 2.10.3 (4 June), and Linux CLI 1.7 (6 April). Versions and promotional prices are volatile; check the live purchase page and release index before buying or installing.

How AdGuard VPN scores in our matrix

The published paid-plan score remains 2.88. We rechecked the stored row on 26 July rather than silently redating it. Its open-source-transparency criterion was already 4/5, so the TrustTunnel release corrects the article but does not change that component or the overall number. The no-logs criterion remains 2/5 because the central claim is still unaudited.

The row's last full verification date remains 21 January 2026 because this desk review did not include new speed, leak or streaming measurements. Calling every criterion fresh would be false precision. The score is a historical matrix snapshot with current documentary corrections, not a re-scored measurement. Inspect the method and compare every provider in the comparison tool.

Who should use AdGuard VPN

AdGuard VPN now makes the most sense for an existing AdGuard user who values simple cross-platform coverage, a modest free tier and an inspectable anti-blocking protocol. The strongest opposing case is that a public codebase plus an app assessment substantially reduces the need to distrust the product. That is fair as far as the client and protocol go.

Choose a provider with a current server-side no-logs audit if your decision turns on proof of non-retention. NordVPN and Proton VPN are disclosed affiliate alternatives with recurring audit evidence; their inclusion does not alter AdGuard's score. The practical recommendation is proportional: use AdGuard for ordinary lawful privacy if its evidence matches your threat model, but do not treat its Google Play badge or open code as proof about production logs.

One useful next step for AdGuard would be a named, scoped, independently published audit of server configuration, retention controls and incident handling, with remediation and a repeat cadence. That would answer the question readers are actually asking.

Frequently asked questions