AntiMalware•July 22, 2026•🇷🇺Translated from Russian

DeepSeek User Conversations Appear in Google Search Results via Publicly Shared Links

Conversations between home users and the Chinese AI service DeepSeek, including many in Russian, have been discovered in Google search results. Accessing another person’s dialogue with the model requires only a specially crafted search query.

The pages in question are Shared Conversations — dialogues that users voluntarily published via public links. Every question, answer, and the titles of any uploaded documents are visible on these pages, as verified by a TASS correspondent. Journalists were unable to download the actual files, even by inspecting the page source code.

Representatives of MWS AI stated that no new data leak has taken place. They noted that search engines had already indexed thousands of publicly shared chatbot dialogues more than a year ago. No reports of a compromise of DeepSeek or unauthorized access to private conversation histories have emerged.

Most likely, the indexed pages are either old conversations or ones that users deliberately made public. It remains impossible to obtain access to a user’s account or private chats through these links. In effect, DeepSeek did not open doors to personal accounts; users themselves left certain conversations publicly accessible, and Google simply catalogued them.

In parallel, Yandex confirmed that links to dialogues with its Alice AI assistant are blocked from search-engine indexing. Such links are created only at the user’s initiative, remain valid for 14 days, and can be revoked at any time. Attached files are never transmitted to the recipient, and continuing another user’s conversation stays private.

Related articles

Habr•Privacy & Surveillance

ONYX 2.0 Removes Central Servers for Fully Decentralized Tor-Based Messaging and Calls

ONYX 2.0-beta marks a major shift from its previously centralized design to a serverless architecture that relies entirely on Tor for all communications. Each user account is now represented solely by a cryptographic key pair generated on first launch, with recovery handled through a 12-word seed phrase. Devices connect directly via onion services, and the application bundles its own Tor daemon on desktop platforms while using tor-android on mobile devices. Voice calls are supported through an embedded local TURN server, delivering acceptable quality with 1-3 second latency despite the anonymity overhead. Message delivery now depends on recipient availability, with undelivered messages stored locally on the sender and retried automatically. The update deliberately removes features such as HTTP or SOCKS5 proxies and link previews to prevent IP leaks or external traffic. Developers released the beta to gather feedback on stability across network changes and multi-device scenarios.

Habr•Privacy & Surveillance

Privacy-Focused Browser Tool Compresses PDFs Locally Without Uploading Sensitive Documents

A developer created a PDF compression tool that runs entirely inside the browser to protect sensitive personal and professional documents from third-party servers. The solution addresses repeated situations where embassy submissions, financial presentations, contracts, and internal reports exceeded size limits, forcing users to choose between installing desktop software or risking data exposure through online services. The tool reduces scanned and photographic PDFs by 66 to 97 percent depending on quality settings while honestly reporting weaker performance on text-heavy files compared with Ghostscript. It works on both desktop and mobile platforms, including iPhone and Android, requires no installation, and continues functioning offline after the first page load. On mobile devices the application respects memory constraints by limiting images to four megapixels, preserving readability for A4 documents intended for printing. The project originated from a single evening script built around Ghostscript and evolved into a full web application after similar compression needs recurred across multiple devices and locations.

Habr•Privacy & Surveillance

Building Prizrak: How a Developer Created a Federated Messenger That Masks All Traffic as Legitimate HTTPS

A developer created Prizrak, a federated messenger with end-to-end encryption where all traffic, including calls, is indistinguishable from ordinary HTTPS connections. The project addresses three common limitations of existing messengers: centralized control points, mandatory phone numbers, and detectable encrypted traffic. It uses real TLS 1.3 handshakes to actual domains, multi-port listening, and a hidden token mechanism inside the encrypted channel. When servers cannot reach each other directly, messages are delivered through a network of storage nodes modeled after Ceph's RADOS system. Voice and video calls run on a native media stack with custom STUN-like functionality and careful UDP buffer sizing to avoid packet truncation. An integrated two-hop VPN reuses the same stealth transport while keeping messenger traffic outside the tunnel.

Habr•Privacy & Surveillance

GrapheneOS Setup Guide: Configuring Pixel Phones for Corporate Surveillance-Free Daily Use

This comprehensive engineering guide explains how to deploy GrapheneOS on supported Google Pixel devices to eliminate corporate telemetry collection. It follows three core principles: rejecting proprietary ecosystems, applying Zero Trust through cryptography and open-source audits, and enforcing strict compartmentalization via isolated user profiles. The tutorial covers official installation via the Web Installer, basic owner profile hardening with PIN shuffling and automatic reboot, and the use of Obtainium for direct FOSS app management from GitHub repositories. Detailed recommendations include privacy-focused tools such as KeePassDX, Aegis Authenticator, AmneziaVPN, Signal, and Fossify applications, along with VPN kill-switch configuration. Regional profiles are created for sandboxed Google Play, Aurora Store, RuStore, and Huawei AppGallery to safely run banking, marketplace, and social apps without cross-profile tracking.