udictio
a social app shaped like a dictionary
9 followers
a social app shaped like a dictionary
9 followers
udictio is a social app shaped like a dictionary. every topic, a song, a movie, a person, a feeling like "loneliness" or "biggest regret", gets one page, and anyone can add their take to it. no feed, no follower counts, no algorithm deciding what you see. entries are chronological, voting just surfaces the standout ones, and vote counts are hidden so an entry stands on what it says. built it solo over two years. iOS for now. would love your honest take.
Free
Launch Team / Built With
Render: Get up to $100K in startup creditsThe fastest way to launch and scale your startup in the cloud
Promoted
so the chronological setup is interesting, but how do you keep people from just nuking entries with downvotes or brigading? is there any moderation layer or rate limiting behind it?
@kader318743 good question. a few layers, some structural, some enforced:
the structure does the heaviest lifting. the default view of any topic is chronological, and vote counts are hidden. so downvotes literally cannot bury an entry in the thread, there's no ranked feed to sink it in. votes only matter on the opt-in "best of" surfaces, and even there the entry never disappears from the topic itself.
on top of that there are hard limits: one vote per entry per account (enforced at the database level), daily vote caps, and a cap on how many times you can ever vote against the same author, which is aimed exactly at the "one user nukes another" pattern.
every entry has a report button, and reports come straight to the moderation panel, moderation is manual right now, suspensions and removals included.
and to your first point, yes, chronological by default (you can flip the sort or filter, e.g. entries with images or from people you follow), no engagement algorithm reordering anything.
genuinely curious how you plan to keep spam and low-effort takes off the pages once the app grows past early adopters, especially without an algorithm or follower counts to lean on
@aleyna798159 thanks for the good question, and honestly the lack of an algorithm helps more than it hurts here. most spam exists to game reach: hijack a feed, farm followers, ride an algorithm. udictio has none of those. a spam entry just sits on one page in chronological order, seen only by whoever opens that topic. there's nothing to exploit, so the economics of spamming it are bad.
new accounts also have a cooldown between entries. every entry has a report button feeding a moderation panel, and there are hard vote caps so scores can't be gamed.
low-effort takes are the harder, more interesting problem, that's culture, not code. the bet is the same one the dictionary-style platforms proved out: the early community sets the writing standard, and the "top" surfaces reward entries that say something. i'd rather grow slower with good writers than fast with noise.
How does moderation work with open entries on sensitive topics like loneliness or regret, especially with voting hidden from view?
@cansev25195 this is actually where the design choices matter most. on sensitive topics, hidden vote counts protect the writer: nobody's raw entry sits there with a visible score under it, and since threads are chronological, a pile of downvotes can't bury or rank-shame anyone. voting exists, but it's private signal, not public judgment.
moderation is the same everywhere: every entry has a report button feeding a moderation panel, review is manual, and hard vote caps stop dogpiling. what i moderate for is abuse and harassment, not whether a take is sad or uncomfortable. honest entries about hard things are exactly what those pages are for.
when entries are chronological, how do you stop a topic page from getting buried or just turning into a wall of low-effort takes after it grows big
@derya1411171 chronological is the default lens, not the only one. every page has a "top" sort where voting quietly surfaces the standouts, plus filters like entries with images or from people you follow. so you pick the lens, instead of an algorithm picking for you.
the wall problem is also literally why i built the on-device apple intelligence summary: once a page collects enough entries, you can generate a summary of what people actually think instead of scrolling all of it.
and low-effort takes mostly don't stack up in the first place: new accounts have cooldowns between entries, and the report button covers the rest. past that it's culture, the early writers set the bar for what an entry looks like.