09.01.2026 00:14 andi (@andi@snactest.sonnenmulde.at) I've been testing #snac2 for a few days now, more specificly I want to see if it can replace my #friendica installation that I have been using since 2019. This will be a somewhat longer piece that is also meant to see how Snac copes with longer Markdown formatted text.
First things first. Friendica is awesome. It is the everything and the kitchen sink of the Fediverse and does a lot of things right. The only downside I see is the reason why I am looking for an alterantive though. Over time it has become harder and harder to host Friendica on my vserver with 8GB of RAM. I could of course opt for a bigger server, but honestly I am only using a fraction of Friendica's features and would prefer something less demanding.
Snac is lightweight, easiy to install, uses no database and does not even require coockies or Javascript in its Web-UI. I mean, how cool ist that!
Compared to Friendica, its featureset is basic at best. But compared to other Fediverse tools like #mastodon it does not fare so bad. But how usable is it as an alternative to my peronal Friendica use.
The good stuff
Setup is really easy for anyone who has self-hosted any services before. The documentation in manpage format might be old school but it is very complete - read it. If you're on Debian you can even use the package from trixie-backports which is kept very current!
It shows the timeline chronologically but with nested threads. This is the only way that makes sense to me and is exactly what I use in Friendica (where different settings are available). This means every new activity on a thread brings it up again (partly collapsed) so you always see the necessary context to follow conversations. For me this is the most important feature, I would not even have continued testing without it. Snac just does it right!
It has the looks! Well at least after applying some stylesheets. Calling the vanilla install basic would already be an overstatement. But have a look at the styles that are linked from snac's Reaedme. I chose the next.css style and have attached an image of how it looks below. The screenshot is from an Android phone and this works remarkably well, even on mobile (some caveats below)!
It's compatible with Mastodon apps. I have only tested #Fedilab because I use it for other stuff too. Everything just works as expected. Although I have to admit, I like the native Web-UI (next.css) even more.
Long Text with Markdown formatting. 500 character per Toot - No thank you, I am not chatting here!
What is missing (or is it?)
There is only the timeline and you can follow Hashtags, that's it. No Channels, no groups, no theme-specific timelines. But honestly, I did not use any of that on Friendica, mainly for performance reasons. So this is a non-issue for me personally.
Following RSS feeds. This feature I will really miss. In Friendica I could direclty follow Feeds and interact with them from my timeline exactly like with ActivityPub Notes or Articles. I will have to use an external translation service or write my own for that (having written an ActivityPub related tool with tootgroup.py before.)
Message formatting is somewhat less flexible but this post shows what is possible still. Sometimes I like to put images in between paragraphs but not having that is also not a big deal.
Event Calendar. This is not a feature I expect Snac to support but I do use it on Friendica and will have to think about how to replace it.
What could be improved
I love that snac works without Javascript and cookies - that's exactly how I still build my webpages. But there is one usecase where a bit of one or the other would make life much easier - when using the Web-UI on mobile platforms.
I have already stated how great the mobile experience is when using the next.css stylesheet. The only downside is that I have to login with username/password again and again, every time the mobile browser is sent to sleep by Android.
Optionally enabling the use of a login cookie would make life much easier. Even better would be enabling Snac to be installed as a Progressive Web App, but I have to admit, I'm not sure how hard or feasible this would be.
Overall, I do not see many shortcomings with Snac. Obviously it offers less features than Friendica, but what it does, it does very well and at least in my view, the important stuff is exactly where it needs to be.
Conclusion
I am pretty impressed with snac. I'll be running my test server for a little bit longer but I do expect replace my Friendica install in the not too distant future.
Adding grunfink@comam.es because maybe you're interested. Please don't treat this as a feature-request or even critizism. It's mostly meant as info for people who have not tried snac yet.
@frasco Essa instância, que roda #Friendica, a atualização é só rodar sudo git pull e a mágica acontece... Como o software tem seu código aberto e é auditado pela "inteligência coletiva" eu me sinto seguro...
Em 90% das vezes só atualizo para versões estáveis (no momento estou usando uma candidata a versão final que é bem estável).
Ou seja, há riscos sim, mas dependendo do serviço, vale a pena correr, em nome da praticidade.
Mit der Lieferung der Sprachdatei ist klar, dass das Friendica-Release unmittelbar bevorsteht. Die neuen Features und Verbesserungen der Usability kann man inzwischen nicht mehr zählen. Daher vorab ein großes Dankeschön an alle Beteiligten.
El frío de esta mañana de enero se siente con fuerza, de esos que obligan a sostener la taza con ambas manos solo por el calor que desprende.
Hoy el plan es simple: avanzar con lo pendiente mientras el ambiente se templa un poco. A veces no hace falta más que un momento de silencio para poner en orden las ideas antes de que el día acelere.
People, especially those in recovery, need to have safe spaces and CWs allow for self moderation (giving the person on the receiving end control over what they allow into their space).
Ironically, Mastodon isn't even particularly good at that. Not by overall Fediverse standards.
How Mastodon does it
Mastodon handles CWs by having them put into a text field that was invented on identi.ca in 2008 as a summary field. Thus, CWs have to come from the poster. And the poster has no other choice but to force the same CWs upon absolutely everyone else, regardless of whether they need that particular CW or not, regardless of what other CWs someone may need. Your posts will be over-CW'd for some and under-CW'd for others.
What's even worse is that this and only this is deeply engrained in Mastodon's culture. From a Mastodon POV, it's completely unimaginable that it could possibly be any different. It's actually being enforced upon the whole Fediverse by Mastodon users.
CWs would be better if they didn't come from the poster. If they were automatically generated on the reader's side, individually for each reader and their needs, based on text filtering.
Mastodon can do that actually. But it's cumbersome. It's as cumbersome as Mastodon's filters altogether: Each keyword needs its own filter defined and adjusted. Next to no-one knows that Mastodon can do that, also because Twitter can't, and who expects Mastodon to have features that Twitter doesn't? Lastly, it was introduced in October, 2022, with Mastodon 4.0, too late for it to become part of Mastodon's culture.
How Hubzilla does it: Filters
Now let's look at Hubzilla. (Friendica, which Hubzilla was indirectly derived from by its own creator, may be similar, but not quite as advanced; @Hella can probably say more about it as a part-time Friendica user.)
For what Mastodon's filters could do before 4.0, Hubzilla only needs one blocklist. Adding a new keyword is as easy as adding a new line to the blocklist. For reader-side CW generation which Mastodon's filters learned in October, 2022, Hubzilla has an optional extra filter list named "NSFW". It has had it from its very beginnings, and it has inherited it from Friendica. Just about everyone in the whole software family agrees: Having your own CWs generated only for you is better than having the same CWs forced upon you as upon everyone else while misusing the summary/abstract field for it.
Hubzilla also has an allowlist for the whole channel. It optionally has one blocklist and one allowlist per contact so you can filter certain things coming from certain contacts and not coming from all the others. It supports regular expressions. And it has a very advanced filter syntax for the blocklists and allowlists that, for example, allows you to filter by whether a message is a repeat (Mastodon: boost), whether it's public or restricted, whether it's a top post or a comment, from whom it came etc. etc.
How Hubzilla does it: Permissions
But self-moderation doesn't stop at filters on Hubzilla. There's also the permissions system, the second-most advanced one in the Fediverse, second only to Hubzilla's own descendants, (streams) and Forte, again created by the same developer. Hubzilla's permissions system works on up to three levels: for the whole channel (Mastodon: account), for each individual contact (Mastodon: follower/followed) and often also for individual content (posts/threads, images/media/other files etc.).
Some examples of you can do with the permissions system:
You can allow only your contacts or only certain ones of your contacts to see your public profile. (In fact, you can make additional profiles and assign them to certain contacts. But both only works with contacts on certain server applications and not with Mastodon contacts.)
You can allow only your contacts or only certain ones of your contacts to see what other contacts you have. (Again, this only works with contacts on certain server applications and not with Mastodon contacts.)
You can allow only your contacts or only certain ones of your contacts or nobody at all to see your previous posts. (Again, this only works with contacts on certain server applications and not with Mastodon contacts.)
You can only allow posts from certain ones of your contacts onto your stream. (This does not affect comments and DMs.)
You can allow only your contacts or certain ones of your contacts or nobody at all to like, dislike and comment on your posts.
You can completely disallow comments on certain posts of yours.
You can only allow your contacts or certain ones of your contacts or nobody at all to send you DMs.
You can only allow
your contacts
certain ones of your contacts
the members of one of your privacy groups (Mastodon: lists on steroids)
a custom combination of privacy groups and certain contacts minus certain other contacts and privacy groups
only yourself
to see and therefore interact with a certain post of yours. This makes DMs actually private.
Also, conversations on Hubzilla aren't bunches of posts loosely tied together by mentions. They're enclosed objects with exactly one post at the top and otherwise comments. All comments always have the same permissions as the top post, and this cannot be changed. This means you cannot reply to a public post with a DM. This also means that, very much unlike on Mastodon, you cannot pull someone else into a DM conversation just by mentioning them. That someone else simply isn't permitted to see anything in the conversation.
There's even more: You don't only own your own posts. You also own all comments on your posts. You and not whoever wrote them. This means you can moderate your own conversations by deleting comments from it. (That is, if someone on Mastodon comments on your post, and you remove it, you do not delete their Mastodon toot from their Mastodon account. You only remove it from your conversation.)
This is real serious self-moderation power.
Hubzilla is entirely built around moderating your own channel and your own stream yourself and, if it is activated, the public stream (Mastodon: local or federated timeline, depending on the admin configuration) along with it. Mastodon's filters work by deleting stuff from your inbox, but letting it onto and keeping it on the local or federated timeline regardless. Hubzilla's filters and permissions work by rejecting it altogether. Content that you reject, and that nobody else on your hub allows in, doesn't enter the hub at all and thus doesn't show up on the pubstream either.
By the way, this also works with forums (groups): You can make a forum private by only allowing its contacts to see its profile. You can make its membership private by only allowing its contacts to see its other contacts. You can hide it from directories. You can make its conversations private by limiting reading permission to only contacts.
How (streams) and Forte do it
As I've said, (streams) and Forte go another bit further.
They don't rely mostly on templates called "roles" to define the permissions of your channel and of your contacts. You can switch individual permissions on and off anytime.
They have an extra switch to let repeats in or not. Hubzilla needs a filter syntax line for that.
And they have the option to make reply control even more advanced: Next to entirely allowing and entirely disallowing comments, you can allow only your own contacts to comment on a certain post. You can generally allow comments on your posts for a certain period of time. And you can allow comments on a certain post only until a certain point in the future.
Pixelfed tem sido o projeto mais escanteiado do protocolo, infelizmente. Não dá para esperar muito.
@arlon, vc teve problema para transferir seus contatos do #Mastodon para o #Friendica também ou teve gente que se perdeu?
Não sei apontar claramente, mas aparentemente houve uma migração total. Tive a impressão de que, sim, parecia que eu não seguia algumas pessoas, porque demoraram muito a aceitar o contato. No mais, foi uma migração bem flúida, não precisei nem baixar um arquivo cvs.
Quanto à migração entre instâncias Mastodon, experimentei fazê-lo em dois momentos: um entre Mastodon-Mastodon, e outra entre Mastodon-Akkoma. Em ambas tive sucesso, mas entre instâncias Mastodon há mais possibilidades de migração, como listas, palavras silenciadas e outras coisas.
A migração Akkoma-Friendica é que não deu certo. Tive de criar um Mastodon para receber os seguidores do Akkoma, para daí passar para o Friendica. Essa é a razão pela qual eu tenho o perfil @arlon!
A migração de contatos entre instâncias do Mastodon ainda não funciona direito? Eu achava que esse problema era coisa do passado. A do Pixelfed não funciona, se a do Mastodon também não funciona direito é uma das maiores promessas do Fediverso que não está sendo cumprida pelas maiores plataformas.
07.01.2026 21:01 display (@display@loma.ml) Ok das Kanäle-Feature auf Friendica ist schon Bockstark. ein paar tags eingegeben, ein paar ausgeschlossen und schon hast du eine timeline nach deinem geschmack #friendica
07.01.2026 16:00 item (@item@hub.netzgemeinde.eu) @SigmundFreudsBartender Und mit Chance trifft man da auf dieselben Friendica- und Hubzilla-Nasen, wie wenn man auf Mastodon bleibt (außer denen, die sich nicht trauen, diaspora* zu aktivieren). Aber auf weniger Schneeflocken.
Ich finde sogar dein diaspora*-Konto in meiner Autovervollständigung (jedenfalls glaube ich, daß du das bist). Now that's what I call Föderation.