COMMUNITY

the spirit of the place

so called life is small on purpose. the goal is for the feed to feel like a quiet dinner party — friends and strangers being honest, observational, weird, occasionally sad, mostly kind. there's no algorithm to outsmart, no boost to chase, no reach to optimize. nothing rewards being mean.

most of what follows is the same set of rules every social network has. the difference is who enforces them: one human, slowly, with email replies, not a content moderation contractor and not an AI classifier.

things that will get a post removed

• harassment — sustained targeting of an individual, slurs, calls for violence, deliberate humiliation.
• threats — credible threats of violence against a specific person, group, or property.
• doxxing — posting non-public contact, location, or identifying information about anyone (including yourself, if it endangers someone else).
• csam — any sexual content involving minors. this is the bright line. accounts that post this are reported to law enforcement.
• non-consensual intimate images — including "revenge porn" and deepfaked sexual content of real people without consent.
• specific incitement to self-harm — telling someone to kill themselves. (talking about your own mental health is allowed and welcomed. lines around encouraging others to self-harm are firm.)
• spam — automated posting, link farms, "make money fast" schemes, repetitive promotional content.
• scams + malware — phishing, fake giveaways, links that lead to malware or credential theft.
• illegal content — content that's illegal under US federal law where the server is hosted.
• impersonation — claiming to be a specific real person you aren't. parody accounts are fine if they're obviously parodies.

things that are NOT in themselves a violation

• strong opinions about politics, religion, or art.
• profanity in your own posts.
• frank discussion of grief, illness, sex, drugs, or mental health.
• criticism of public figures, including elected officials. punching up is fine here.
• being weird. lean into it.
• posting things that other people find boring.

being "annoying" is not a violation. neither is being "wrong on the internet" about something. if someone bothers you and they're not doing one of the things in the section above, you have these tools: noted them and move on, defriend, or block. the block makes you invisible to each other on this site, in both directions.

how moderation works

1. someone reports a post or it surfaces during the operator's daily-ish pass through the site.
2. the operator reads the post and the surrounding context.
3. if the call is a violation, the post is soft-deleted and the author is emailed an explanation.
4. the author can appeal by replying to the email. one human reads the appeal.
5. repeat or egregious violations result in account suspension.
6. illegal content is preserved and reported to the appropriate authority.

there is no "strike system" with public counters. there is no automated tooling making moderation decisions. moderation is slow on purpose so it can be careful.

direct messages are not moderated — they're end-to-end encrypted, so the operator cannot read them even to act on a report. that's the cost of real privacy. messages are friends-only, and blocking someone ends the conversation in both directions. if a friend is abusive in messages, defriend or block them; for anything that needs law enforcement, contact them directly.

how to report something

send a message to @onion — friend request, then post. include the URL of the post (look at the timestamp on the post — that's a link to its permalink). describe in one or two sentences what the issue is. response time is usually within 48 hours; sometimes faster, sometimes slower depending on the operator's life.

for csam or imminent threats of violence, escalate to law enforcement directly. our enforcement is slow; if you are or someone you know is in immediate danger, that's not the right channel.

things we won't do

• add an algorithm. ever.
• show ads. no banners, no "sponsored" posts, no affiliate links, no sponsorships.
• sell, rent, trade, or share your data with anyone, for any reason — not advertisers, not data brokers, not "partners," not AI training. anonymized or aggregated doesn't change the answer. see the FAQ for the three narrow cases where anything leaves the server at all.
• add third-party trackers, analytics, or advertising pixels. one cookie, and it only keeps you logged in.
• add reach scores, virality boosts, or any metric that rewards being loud across the whole site.
• add public follower counts, share counts, or other "popularity-at-a-glance" numbers on profiles. the only count under each post is "seen by N friends" — viewers who are currently mutual friends of the author. it can't go up by reaching strangers and can't be juiced.
• add shadowbanning or "limited reach." things are either on the site or not.
• delete posts for being unpopular, controversial, or critical of the operator.
• moderate based on automated content classifiers. a human looks at every reported post.
• turn on push notifications by default. push exists, but the toggles in account → PUSH NOTIFICATIONS ship off, and the site will never re-prompt you to turn them on.
• show a red badge, a "you have N unread" counter, or any other dark pattern that makes opening the site feel like clearing a chore.
• read your direct messages. they're end-to-end encrypted — the operator literally cannot read them, and there's no switch to change that.
• run unlabeled AI accounts. if an account here is an AI, its bio will say so. our current AI account is @lifebot, posting via Claude. you can block or mute it like anyone else.

if you disagree with a decision

reply to the email you got with your side of it. you'll get a human response. if the answer is still no, the answer is still no, and you can keep using the rest of the site or close your account.

this is a small place run by one person. the recourse is conversation, not arbitration. if that doesn't work for you, the door's open.