Freenet/Hyphanet - A mix of USENET, IPFS, BitTorrent, I2P, and TOR with user-moderated uncensorable forums!

[CRkVhG2pNFI1Qmzv5MXJfe]

When examining and discussing here, I ask you to pay special attention to FMS, a pseudonymous and truly decentralized forum system on Freenet. Anyone can say anything on any board. Each user gets to decide whether any other user should be there, and whether they trust other users' judgements on this question. This forms a web of trust that enables effective moderation without administrators or moderators on an uncensorable forum. To me this makes it perfect as a fallback during outages of this site. What do you think?

Quick search showed Freenet sometimes mentioned, but no dedicated thread. It seems extremely useful and unique, despite predating BitTorrent, TOR, and I2P. Unlike other anonymity services, you do not need to host/seed content yourself. The network stores your content for you. Every user contributes encrypted block storage (not the entire file like in BitTorrent) and bandwidth (not just a few anointed nodes like in TOR). Has update-able file and site keys like IPFS.

Hyphanet is the original Freenet. There is a web3-focused reboot called Freenet/Locutus which drops privacy from the list of requirements for faster speed, and is not the topic of this thread. From the Hyphanet project website:

I worry about my child and the Internet all the time, even though she's too young to have logged on yet. Here's what I worry about. I worry that 10 or 15 years from now, she will come to me and say 'Daddy, where were you when they took freedom of the press away from the Internet?

Spoiler: local archive

Hyphanet is free software which lets you anonymously share files, browse and publish "freesites" (web sites accessible only through Hyphanet) and chat on forums, without fear of censorship. Hyphanet is decentralised to make it less vulnerable to attack, and if used in "darknet" mode, where users only connect to their friends, is very difficult to detect.


Communications by Hyphanet nodes are encrypted and are routed through other nodes to make it extremely difficult to determine who is requesting the information and what its content is.


Users contribute to the network by giving bandwidth and a portion of their hard drive (called the "data store") for storing files. Files are automatically kept or deleted depending on how popular they are, with the least popular being discarded to make way for newer or more popular content. Files are encrypted, so generally the user cannot easily discover what is in his datastore, and hopefully can't be held accountable for it. Chat forums, websites, and search functionality, are all built on top of this distributed data store.


Hyphanet has been downloaded over 2 million times since the project started, and used for the distribution of censored information all over the world including countries such as China and in the Middle East. Ideas and concepts pioneered in Hyphanet have had a significant impact in the academic world. Our 2000 paper "Freenet: A Distributed Anonymous Information Storage and Retrieval System" was the most cited computer science paper of 2000 according to Citeseer, and Hyphanet has also inspired papers in the worlds of law and philosophy. Ian Clarke, Hyphanet's creator was selected as one of the top 100 innovators of 2003 by MIT's Technology Review magazine.


An important recent development, which very few other networks have, is the "darknet": By only connecting to people they trust, users can greatly reduce their vulnerability, and yet still connect to a global network through their friends' friends' friends and so on. This enables people to use Hyphanet even in places where Hyphanet may be illegal, makes it very difficult for governments to block it, and does not rely on tunneling to the "free world".

 

[CashMoney!]

This looks interesting. Thanks for sharing. I think the concept of "digital gardens" like this is where we need to go. You have your own corner that isn't constantly intermingled with content from algorithms for profit and addiction. You choose your level of engagement and visibility. I agree with the guy in the video so far that NYT being bigger than Doom is ridiculous and the modern web is not built for the users anymore. Excited to try this and contribute to the project.

 

[SleepingBull]

I've dipped my toes into Freenet a few times over the years. As far as I'm aware it maintains a userbase of around 10k users, a creepily common population I see for many obscure systems.

The moderation system of FMS, which I always thought of simply as Frost, as that was the common UI you'd use to access it, is based on you setting trust ratings of others and inheriting some amount of the trust ratings from those you trust when it comes to people you haven't rated personally. In some ways it can act like an opt-in moderation system where you can set high trust to a small group to make them essentially act as your moderators. There are a lot of options to change your own configuration of the system, and most of it is pointless.

In theory, the idea is that spammers would be assigned -10 trust by everyone who saw them, and if you have a few humans you trust all those -10s would result in the spam's score being so low your client would block it. In practice: Spam never really developed on the FMS, so the whole system is ultimately unnecessary and you can just set your client to show every message with more than -10000 trust. There are a couple of competing groups of trust that distrust each other on FMS, or there was when I last looked at it half a decade ago. The usual darknet divide of CP versus everything else. I wonder if things might get weird if there ever was seriously sized communities on the system, as various forum spats might be prone to turning into groups negrating each other and unknown bystanders suddenly finding half the forum disappearing because they had mostly only posrated one group and was now subscribed to their ratings hard enough to accidentally take sides. All of which is to say that the system which was supposed to be "Rate this person up if they're human, down if they're spam" would inevitably get hijacked for a different purpose. Still, it is the best solution to the problem I've seen, not that I've actually seen many attempts to solve it.

 

[Kosher Dill]

FMS and all other "decentralized forums" invented so far seem to throw out just about all the aspects of a forum that make it worthwhile in the first place. It's an inherently centralized undertaking, and coming up with a less-centralized application that preserves the useful features would probably just look like what Null set up to keep this place running.

 

[AmpleApricots]

In theory these things always sound great but in practice you visit these places and it's just pedophiles, schizos, political extremists and trolls (not rarely a combination of several in one person) all proud and loud about their disgusting existence because they know there's no way to truly remove them. And across all these platforms, I'd swear to god it's always the same group of people. I have no proof for this but they always behave extremely similar. These absolute dregs of society even started creeping in in times where kiwifarms was mostly limited to tor. What (mostly, not always) drives them away when kiwifarms is reachable via clearnet I'll never know but it seems to somehow work that way.

 

[sweater_supremacy]

as much as this seems like a nice concept i am doubtful that it would work as anything other than a safety net in practice, when the clearnet site is down and only the .onion link is available active users tank on here

 

[i440BX]

This seems promissing.
In the worst (improbable) case of making internet fragmented, kyc-ed, this might be the solution.
Or we might develop low speed networks like people in Cuba that are able to play network games. Or we might start trading usb-sticks/DVDs.

 

[CRkVhG2pNFI1Qmzv5MXJfe]

FMS, which I always thought of simply as Frost

I think Frost is FMS predecessor. Frost has no web of trust and can therefore be DOSed easily by a single individual. FMS is its successor. It is still working as intended.

[forums are] an inherently centralized undertaking

That is new to me. Why would a community-moderated forum be any less of a forum?

In theory these things always sound great but in practice you visit these places and it's just pedophiles, schizos, political extremists and trolls (not rarely a combination of several in one person) all proud and loud about their disgusting existence because they know there's no way to truly remove them.

This due to technical illiteracy among normals. Same goes for TOR and this site. When we are on TOR, the site is still the same, but it is much less popular and more schizo. Not TOR fault.

as much as this seems like a nice concept i am doubtful that it would work as anything other than a safety net in practice, when the clearnet site is down and only the .onion link is available active users tank on here

What can be done to alleviate user burden?

Or we might start trading usb-sticks/DVDs

Freenet supports sneakernet transmission.

 

[sweater_supremacy]

What can be done to alleviate user burden?

i honestly think that having some form of clearnet access is the only solution for that, people like using kiwifarms but a good chunk of them probably arent willing to resort to third party services, a lot of people didnt even bother using tor to access the .onion link despite the fact tor's onionizing feature would automatically redirect them to the onion version if they typed in the clearnet adress, the best way to alleviate the burden is seemingly to just make sure theres a way to access the site from the same browser that is being used as the daily driver

 

[Kosher Dill]

That is new to me. Why would a community-moderated forum be any less of a forum?

Look how hard Null & co. have to work to keep us autistic retards in line here. Do you really think it would be the same if the retards were allowed to run the asylum? No, I think traditional, centralized moderation (and identity management) is a core part of what makes forums more than just imageboards with threading. You have to be able to set ground rules and a single, consensus view of what's "in" and what's "out".
Alternatively, if you don't care about that stuff, you don't need a forum at all. Just have an imageboard, or a Usenet group, or a group CC-all email chain with threading enabled in your email client.

 

[y a t s]

I feel like I'm being sold something. Call me a schizo, call me retarded, but I get the same gut feeling I do while reading an MSM article about "How XYZ is Actually a Good Thing".



Regardless, this is an interesting idea, but Freenet was full of glowies last I checked.

 

[soup peddler]

Freenet is 10% schizos, 85% pedophiles, 5% cybersecurity people who may or may not fall into one of the aforementioned groups.

 

[The Mass Shooter Ron Soye]

https://megalodon.jp/2026-0208-2348-37/https://www.youtube.com:443/watch?v=3SxNBz1VTE0

December 25, 2025: River: Real-Time Group Chat on Freenet (archive)

February 8, 2026 - Freenet v0.1.114 Released

 

[PlugsPugs]

I feel like I'm being sold something. Call me a schizo, call me retarded, but I get the same gut feeling I do while reading an MSM article about "How XYZ is Actually a Good Thing".

Ver archivo adjunto 5696787

Regardless, this is an interesting idea, but Freenet was full of glowies last I checked.

That there is a database of anything, tells me hypha net is compromised. And that analysis states specifically that there is a database of objectionable materials, and an ability to determine who has those chunks of data.

Of course, I detest the content tracked. However I assume any tech used to control "the bad stuff" will inevitably be used to control the "not bad stuff", the "inconvenient for the wrong people stuff", and the "wrong-think stuff".

 

[y a t s]

That there is a database of anything, tells me hypha net is compromised. And that analysis states specifically that there is a database of objectionable materials, and an ability to determine who has those chunks of data.

Of course, I detest the content tracked. However I assume any tech used to control "the bad stuff" will inevitably be used to control the "not bad stuff", the "inconvenient for the wrong people stuff", and the "wrong-think stuff".

What gets me is that there is basically nothing stopping an agent who barely passed STATS101 from piecing together where file requests originate. It's child's play (no pun intended):

[16] When a user wants to retrieve or access a file from Freenet, the user must possess a unique identifier for that file, known as a manifest key. The manifest key is a long alpha-numeric string. To download a file, the user inputs the manifest key into Freenet’s user interface and hits a download button. The user’s computer then sends out requests for the file. Several requests are sent for any given file. Those requests each contain a “split key” which is a unique identification key for one block of a file.

[17] The requests begin to travel the diverse network of nodes that make up Freenet, searching for blocks of the desired file. When a node receives a request, it checks its own data store to see if it has the block associated with the split key in the request. If the node does not have the requested block, it forwards the request to another Freenet node, which again checks its data store. The request is routed among various nodes until it finds a node that has the blocks. If a Freenet node has the requested block in its data store, it will send the requested block directly back to the Freenet node that requested it following the same path by which it received the request. Once the original requesting node has received all of its pieces, it will stop sending out requests for any pieces of the file.

[18] When a police officer wants to determine if a Freenet node is downloading child pornography, he parses the logs kept by the ICAC database. That database logs activity on Freenet from several law enforcement nodes operating on the network that have been modified for law enforcement use. In brief, it keeps track of blocks of known child pornography files passing through those law enforcement nodes.

[19] As mentioned above, a request for pieces of a file may travel through several nodes on its way to finding a node with pieces of a file. In other words, there might be many “innocent” nodes along the path of a request that are simply passing on the request to other nodes. The investigator’s job is to determine which nodes are actually requesting the child pornography files, as opposed to simply re-routing them. The re-routing nodes are merely passing along a request and their users are not interested in downloading the file. Only the Freenet user at the requesting node wants to download the requested file.

[18] is interesting because they don't inherently need to store such content (in the traditional sense) and allow others to download it. All they have to do is pass data down the chain and log anything they deem suspicious.

[24] The applicant argued that one statement in the ITOs is so fraudulent that it engages this court’s residual discretion to set aside the orders. The following statement appears in the ITOs for both investigative orders as part of the background information about how Freenet works:

There is a counter on this process of repeated requests, which will decrease each time the request is past [sic] from node to node. The default number is 18, meaning the request should be past [sic] no more than 18 times before it stops looking for that piece.​

[25] When he was cross-examined, the affiant acknowledged that this information was not complete. Freenet does use a counter. When a request is sent, it will not endlessly bounce around the network requesting pieces of a file. The request will only “hop” between nodes searching for the request a set number of times after which the request will expire. The number of times a request will bounce around among nodes is known as “hops to live.” The default number of hops to live is set at 18. The affiant explained this much correctly in the ITO. However, he left out an important feature about how this counter works.

[26] As the affiant explained in cross-examination, when a Freenet connection is first established, the first node to which a request is sent randomly determines whether to have the counter decrement by one (to 17) or remain at 18. After that first “hop,” the counter decrements normally by one for each “hop.” This feature is meant to obscure the identity of the requesting node because it prevents an accurate backwards count to the requesting node by simply counting back the number of hops on the counter.

The feature mentioned in [26] is a laughably shitty trick because it only cuts the odds in half that a request with 18 remaining hops was sent from the originator (i.e. it's either the originator or the node 1 hop down from him). With basic correlation techniques, this 50% split becomes essentially useless.

[45] In his explanation about why he believed that the three particular files had been downloaded by the suspect, the affiant explained in the ITO what he had observed from his review of the logs. The following is his explanation about the file downloaded on December 15, 2014:

ii. This database had had 132 pieces requested out of 606 total pieces of the file needed. Some of these are re-routing for other users, but I believe many are a direct result of a Freenet user, the suspect, downloading the file.​

​

[iii. contains the manifest key for the file]​

​

[iv. is blank]​

​

v. Volume of requests: I looked at the individual requests and there were no requests that looked to be not consistent with this node requesting the file. This is 132 pieces of the file, which is 21.78% of the file. I know this to be above the threshold needed to believe it is a requester.​

​

vi. Time: This download took 23 minutes to collect all of the pieces from the Freenet network. It started on the 15th of December 2014, at approximately 4:13 am (UTC) and finished at 4:37 am (UTC). I know this to be consistent with normal operation of a Freenet download. vii. There were also multiple instances that more than one request was coming in per second. This is consistent with this target node being a requester of the file, not re-routing a request.​

​

viii. Even share: This suspect Freenet node averaged 11.2 peers during the download time. The suspect node will try to evenly spread out the requests across these peers, which would be at approximately 54 requests each. In this case, there were 132 requests. This is consistent with this target being a requester, not re-routing.​

​

ix. Overall, I have analysed the Freenet logs in relation to this IP address and suspected child pornography file. I believe that this IP address was requesting the file, not simply rerouting a request from a different Freenet user. If that were the case, I would see less total requests in both quantity and even share. The time is also important, having multiple records coming in within a second. Re-routing has some processing time at each node and takes longer.​

 

[i440BX]

From the looks of it, I2P seems more fit for the purpose than hyphanet.

i honestly think that having some form of clearnet access is the only solution for that, people like using kiwifarms but a good chunk of them probably arent willing to resort to third party services, a lot of people didnt even bother using tor to access the .onion link despite the fact tor's onionizing feature would automatically redirect them to the onion version if they typed in the clearnet adress, the best way to alleviate the burden is seemingly to just make sure theres a way to access the site from the same browser that is being used as the daily driver

I think the easiest way for people to use tor is to switch to Brave.