Freenet/Hyphanet: The Forgotten Sibling of the Darknet Family

Freenet/Hyphanet: The Forgotten Sibling of the Darknet Family

So you think you’re anonymous because you’re on Tor, huh? That onion in your URL bar makes you feel like a ghost in the machine, a digital ninja slipping through the cracks of surveillance capitalism. And yeah, you’re not wrong, Tor is great. But let me take you on a journey to meet its weird, brilliant, slightly neglected older sibling: Freenet, now known as Hyphanet.

No, it’s not a new altcoin. It’s not a web3 meme project. And it’s definitely not a distributed AI running on Raspberry Pis in your neighbor’s basement (…yet). It’s a darknet, like Tor. But unlike Tor, which acts like an anonymous proxy for accessing content, Freenet is a decentralized, anonymous data store, a place where content lives, not just a tunnel to get to it.

Let’s plug in the modem, set our clocks to the late ‘90s, and dive into the tangled, occasionally brilliant, often misunderstood web of Freenet.

A Seed is Planted: Freenet in the Beginning

The Freenet Project was born in 1999, back when Google was a scrappy little startup, Napster was pissing off Metallica, and the biggest privacy threat was someone reading your ICQ logs.

Enter Ian Clarke, an Irish computer science student with a rather spicy final-year project: a decentralized, censorship-resistant publishing platform. While his university wasn’t exactly thrilled (they gave him a B), the idea struck a chord with the early internet freedom crowd. He posted his paper online, titled “A Distributed Decentralized Information Storage and Retrieval System.” Sexy, right?

The idea was simple but radical:

• No central servers.
• No domain names.
• No way for authorities to know who published what.
• Data lives on participating nodes, encrypted and chopped into little chunks.
• The network routes around damage, censorship, and even user stupidity.

He called it Freenet, not because it was about free software (although it was), but because it was about freedom of speech, thought, and information.

Early Days: Pirates, Rebels, and Anarchists

In the early 2000s, Freenet attracted the kind of crowd you’d expect: dissidents, whistleblowers, digital anarchists, and of course, the usual internet weirdos.

• Want to publish political dissent from behind the Great Firewall of China? Freenet.
• Need to share documents proving corporate malfeasance? Freenet.
• Want to distribute controversial literature banned in your country? Freenet.

It quickly became known as a place for the uncensorable web. The concept of “insert once, read forever” was both its power and its curse. Once a file made it onto Freenet, it stayed there, indefinitely, redundantly, and anonymously, as long as someone requested it occasionally.

This also meant that questionable content, you know, that kind, could potentially persist. Freenet developers implemented filters and content-hiding features later, but the stain of “anything goes” stuck, and it scared off a lot of normies.

Meanwhile, over in onion-land, Tor was busy becoming the darling of journalists, human rights activists, and the EFF. Freenet, meanwhile, was mutating in its dark little cave.

Enter Darknet Mode: Trust Your Friends, Not Strangers

Originally, Freenet was a “global” network, you connected to random peers and became part of the whole swarm. But this opened up attack vectors. A well-resourced adversary (say, a government) could flood the network with nodes and de-anonymize people.

So the Freenet team introduced “Darknet mode.” No, not that darknet, I mean the Freenet flavor. In this mode, you only connect to people you explicitly trust, your real-life friends, your anarchist book club, your cat who runs a node on the side.

This makes surveillance way harder. Instead of being part of a huge flat network, you’re a branch on a weird encrypted social tree. If the police want to catch you publishing banned literature, they have to infiltrate your actual social circle. Harder than it sounds when your friends are paranoid nerds who think signal isn’t encrypted enough.

This peer-trust model inspired later projects like RetroShare and GNUnet, and it made Freenet feel more like a secret society than a regular network.

The Content That Refused to Die

Freenet’s storage model is weird. It’s persistence through popularity. If nobody asks for a file, it disappears, overwritten by newer data. But if people keep requesting it, it sticks around, replicated across nodes like digital kudzu.

There are three main key types:

• CHK (Content Hash Keys): Immutable. You can’t change the file without changing the key.
• SSK (Signed Subspace Keys): A way to create updatable “sites” and pseudonyms.
• USK (Updateable Subspace Keys): Like SSKs, but with versioning.

This gave birth to the idea of Freesites, static websites that lived inside Freenet, loaded block-by-block from the network. You could browse forums, blogs, or even file archives, no DNS, no IPs, no servers. Just keys and data.

Yes, Freenet had decentralized web pages long before IPFS made it cool.

Mid‑Era Evolution: From 0.5 to 0.7 — Mind the Gap

After the glorious mess of the early 2000s, the Freenet devs took a long look at version 0.5, sighed deeply, and collectively muttered: “Yeah… this is gonna need a rewrite.”

And thus began the road to Freenet 0.7, a massive overhaul launched in 2008, where practically everything got re-engineered: routing, connectivity, anonymity layers, you name it. But wait… 0.6? Never heard of her.

There never was a 0.6. Just like Netscape skipped version 5 (because Mozilla was busy doing actual work), and Microsoft jumped from Windows 8.1 straight to 10 (presumably to pretend 9 never happened), Freenet just… yeeted over 0.6.

But why 0.7 Was a Big Deal

Version 0.7 wasn’t just a patch—it was a philosophical shift:

• Freenet 0.5 relied on dynamic routing with automatic peer selection. That was great for discovery… and for surveillance.
• 0.7 introduced darknet mode: you connect only to people you trust, forming a social-web-of-trust that’s massively harder to map from the outside.
• They also introduced hybrid mode, letting you mix in untrusted “opennet” peers if you didn’t have enough darknet contacts.
• Under the hood, the devs ditched clunky TCP for leaner, NAT-punching UDP.

Result: the network became faster, more anonymous, and much more resistant to sybil and analysis attacks.

It was still written in Java (of course), but with an evolving plugin system, better UI via FProxy, and community-developed tools like Sone and FMS. By mid-2009, version 0.7.5 was rolling out with huge memory improvements, smoother content caching, and more stable darknet handling.

In short: Freenet 0.7 made it past the “cool idea” phase and into the “actually usable, but still very nerdy” category.

The Rebrand: From Freenet to Hyphanet

In 2022, something weird happened.

A new project popped up on the web under the name Freenet, but not that Freenet. This one had a shiny modern website, vague gestures toward decentralization, a splash of blockchain-flavored marketing lingo, and exactly zero of the original source code or principles. It was like someone spray-painted “Freenet” on a rented scooter and tried to pass it off as a privacy tool.

That someone? Ian Clarke, the original creator of Freenet back in 2000.

Yes, the same Ian Clarke who’d long since vanished from the dev community, only to reappear two decades later, rename his new (and unrelated) project Locutus to “Freenet,” and proceed to commandeer the name, website, and branding, all while ignoring the very community that had kept Freenet alive for years without him.

No merge. No Rust rewrite. No outreach. Just a cold reboot of the brand, with a heavy dose of venture-scented buzzword salad and a complete abandonment of Freenet’s original privacy-first philosophy.

The feeling of betrayal was most hard felt because what is now called Freenet holds few, if any, of the core principles or goals of the pre-existing network. If it HAD just been a re-write in Rust, then most core devs and users would have probably been happy to jump on board. Instead we get a buzzword salad cash grab of a project that leans heavily on the brand name earned through hard work and community built over 2 decades, which puts zero priority on anonymity or privacy…

Needless to say, this didn’t sit well with the actual Freenet developers — the ones who’d been maintaining the code, fixing routing issues, moderating forums, and building tools since forever. To them, it felt like being evicted from your own squat by a guy who sold the front door to a startup.

Enter Hyphanet

Rather than fight over the name, the core team made a bold — and frankly classy — decision: they rebranded themselves.

Thus was born Hyphanet, named after hyphae — the underground filaments of fungi that quietly sustain vast ecosystems. A poetic metaphor for a network that thrives in the dark, survives by trust, and spreads organically.

Hyphanet is, in every technical and ethical sense, the true successor to Freenet:

• It runs the original codebase.
• It maintains the original goals: strong anonymity, censorship resistance, and user empowerment.
• And most importantly, it’s still developed and maintained by the same people who didn’t abandon the community for 15 years.

So while the name “Freenet” may now point to something unrecognizable, the real thing never died — it just changed clothes and grew stronger roots.

Present Day: Hyphanet in 2025

So what shape is the original Freenet in as Hyphanet today, mid‑2025?

Codebase & Releases

Hyphanet continues the Java 0.7.x heritage, hosted under the hyphanet/fred repository. The latest builds (since 14 98) finalize Windows signing, installers, and incremental UX updates. The project still ships the same mix of darknet/opennet, FProxy, plugin architecture, and media support.

Community & Ecosystem

• Plugins like Sone (social), FMS (forums), FLIP (chat), and media streaming keep evolving.
• Localization spans many languages, German, Persian, Japanese, Swedish, Russian, etc., reflecting global participation.
• Community infrastructure lives across IRC, GitHub, and identity-rich peer forums.

In the Real World

• Not massive adoption, still niche, finicky, and best‑used by enthusiasts, but it remains the canonical peer‑to‑peer datastore with deep roots.
• Its darknet mode provides plausible deniability, but the architecture has raised concerns: nodes cache encrypted data without knowing its content, meaning users could theoretically host something illegal, without ever being aware of it.

Road Ahead

The biggest hope for Hyphanet now is gradual modernization, not flashy relaunches. They’re focused on:

• Fixing installer and UX quirks
• Enhancing darknet robustness
• Improving plugin ecosystem
• Keeping infrastructure solid and maintainable

Conclusion: Why you should join?

By the time Clarke decided to hijack the name in 2023 (remember that betrayal chapter?), Hyphanet had already solidified its identity, as the original decentralized, anonymous datastore. It’s not shiny. It’s not crazy fast. It’s not huge.

But if you value data survival, resistance to censorship, and peer‑to‑peer civilization, Hyphanet is one of the few systems that actually delivers, and has done so for nearly two decades.