The Anonymity Networks Community is now accepting proposals for the COSCUP 2026 Anonymity Networks Community Track. This is our second year running related sessions at COSCUP. Through this open CFP, we hope to invite more people who care about or actively practice anonymity to share their work. The track runs for two days and includes talks, workshops, demos, and field experience sharing.
This post is based on the Tor Project announcement:
Arti is Tor’s next-generation Rust implementation. The 2.2.0 release is notable because it pushes a previously experimental access path into day-to-day usability: HTTP CONNECT is now included in full builds and enabled by default, sharing the same port as SOCKS.
For teams operating in filtered enterprise, campus, or public networks, that matters immediately. For developers embedding Arti, this release also expands RPC ergonomics with non-blocking requests, event-loop integration, and a new superuser administration path. In one version, Arti improves both network practicality and operational controllability.
We're excited to share that the Tor Project invited us to contribute a guest blog post about our experience running a Tor relay on a university network in Taiwan. You can read the full article here: Setting up a Tor Relay at a university in Taiwan.
Taiwan occupies a unique position in the global internet freedom landscape. While the country enjoys relatively open access to the web, it operates under persistent geopolitical pressure and is regularly targeted by sophisticated cyber operations. In this context, privacy tools like Tor aren't fringe utilities — they're practical infrastructure for journalists, researchers, civil society organizations, and anyone who needs to communicate or organize without being observed. Building awareness and local capacity around these tools is part of what our community is working toward.
We follow Tails releases because they ship the same building blocks many of us recommend in real life: Tor, a hardened desktop, and tools for people who cannot assume a “normal” network path. 7.6, dated 2026-03-26, is worth translating not for one killer feature, but for two changes that affect how people get online and how they store secrets on a live system.
Tor bridges are not exotic; they are often the difference between “Tor works” and “Tor never connects.” In places where Tor traffic is filtered or throttled, users learn to hunt for bridges through side channels—paste sites, trusted contacts, or ad‑hoc instructions. Tails 7.6 brings that guidance into the Tor Connection assistant: pick Connect to Tor automatically, and if the network blocks Tor outright, the bridge screen can Ask for a Tor bridge based on your region, pulling candidates via the Tor Project’s Moat service—the same family of tech Tor Browser has used since 11.5—with the fetch disguised using domain fronting.
For readers in Taiwan and across East/Southeast Asia: censorship models differ, but the pattern is familiar—TLS interception, routing games, or “soft” blocking that fails open only for some apps. A Tails image that surfaces bridge acquisition in-product lowers the bar for journalists, lawyers, and civil‑society volunteers who already juggle operational risk; they should not also have to memorize bridge workflows from blog posts.
The second headline is Secrets replacing KeePassXC. That is a product decision, not a security downgrade by default: Secrets is tighter with GNOME, which matters on Tails because accessibility regressions (on‑screen keyboard, cursor sizing) are real blockers for some users. KeePassXC power users can still add it via Additional Software; the database format overlaps, so migration is meant to be frictionless.
In February 2026, anarcat from the Tor system administration team (TPA) published a post titled \"Keeping track of decisions using the ADR model\".
After reading it, we felt it offered a very practical way to think about proposals, decision-making, and how to write things down so that people can actually find and understand them later.
This post is not a translation of the original article. Instead, it is our own summary and reflection on:
On 2026/02/09, Discord announced a global “teen-by-default” rollout and stronger age verification (English coverage: BBC, Medianama). New and existing users will default to a teen-oriented experience; to relax content filters or access age-gated spaces, users must complete verification via facial age estimation or by submitting ID. Discord frames this as a commitment to teen safety and Safer Internet Day, and will use an “age inference model” in the background to help assign age groups.
We are not dismissing Discord’s intent—youth protection and compliance are serious. But such measures also mean one thing: large platforms will need more personal data and behavioural signals to “classify” users. Whether via face scans, ID documents, or algorithmic inference, the result is handing over “who you are, how old you are, where you are” to the platform and its partners. For many people who just want to chat, game, or collaborate, that may be an acceptable trade-off; for others, it raises the question: is there another way?
Commercial chat platforms have their own rules: terms of service, product direction, what data is retained, how algorithms and policies work—mostly driven by the company and shareholders, with little say for ordinary users and little visibility into how their data is used. This isn’t about “who is worse”; it’s about who gets to decide.
The anoni.net community has chosen a different path: self-hosting a Matrix homeserver. We run tuwunel, a high-performance Matrix homeserver implemented in Rust, on im.anoni.net for community discussion and 2026 theme collaboration. Server configuration, retention policy, and channel rules are decided by operators and the community together—smaller scope, more predictable, and more transparent. Our focus is clear: internet freedom, anonymous networks, and privacy in practice, not “anyone can join and talk about anything.” This is a themed, consensus-oriented workspace.
This time, we’re heading south to Kaohsiung (a.k.a. Takao) to participate in the “g0v Hackath71n – g0v the 71st Hackathon.” If you happen to be in southern Taiwan during this period, or if you’d like to travel south to Kaohsiung together—combining a short trip with a few days of remote work while joining the event—feel free to come find us at the hackathon. We’ve registered a booth to help move forward the progress of the “Anonymous Online Community” 2026 project!
At the moment, it looks like there are still a few spots available. If you’re interested in our project and would like to contribute, you’re very welcome to join us on the day of the event—no matter your background or area of expertise.
As 2025 comes to a close, we sincerely thank all our partners for their participation and support throughout the year. From project-based efforts to community building, we have accumulated many actions and valuable experiences over the past year. Below, we will look back on the key initiatives of 2025, and also share our next steps for promoting anonymous networks in 2026.
In 2025, the international human rights–focused conference RightsCon was held in Taipei. Upon learning that the Tor/Tails and OONI teams would also be visiting Taiwan, we initiated and prepared related workshop events several months in advance. The event attracted over 300 participants and specifically invited partners from news media and civil society organizations to join the discussions.
After the event, we compiled and published a comprehensive retrospective article to document this rare and valuable exchange.
As a result of hosting this event, we also recruited many enthusiastic volunteers. Together with them, we continued subsequent preparations, laying the groundwork for related activities held in August 2025.