Post

Lit! But too much info for average human brain and too fast to average human eyes 😅 probably need another layer to filter the noises for average human beans… Unfortunately I’m not a dev :(

you can hit the filter button and do some exclusions or include certain keywords/people only. Actually just turning off reply posts cuts things down pretty well. Skeeters love to reply

Going back in time is tricky, would require storing all posts - I am thinking about offering persistent queries at some point, ie you could close/reopen the tab for a narrow query and be able to see what you've missed

it's spiritually a monday lmao it would be helpful if you could open-source at some point tho! I couldn't figure out how to get profile images from the firehose, maybe I just didn't look close enough

This feed could really help tune the What's Hot! tab since I was able to find the Mark Cuban AMA happening in real time via this firehose, when it wasn't showing up in tab. I noticed that you included the follower count in parentheses, & it'd be nice to filter by max followers or min followers.

i could do it, but bear in mind this is live stuff - before any labels can be applied, even automatically. could be extremely nsfw, you'd be seeing the raw id of the sky-o-verse maybe i'll add an optional switch with a big warning label

yes you can already id some spammy bot type accounts that just post nonsense over and over, i could maybe add them to a default block list, and have custom muting per user browser the little number next to the user handle name is their follower count

would love to, but how do i know the second after it's posted? this is very real-time, pre labels or anything. image how many ways this could go wrong maybe if i could get an image classifier working in the browser locally... not easy

a thought you may have already had: currently, filtered words are partial matches ("the" matching "theater"); doing a word boundary regex search might make more sense?e.g.: part.text.toLowerCase().includes(token.lowercasedText); to: .match("\\b"+token.lowercasedText+"\\b") in `meets`?

Cool! Would nice to have this for all the likes flying around, not sure there is a way to pull those from the API yet. Then I could filter on incoming likes to my posts, see the pfp of likers and the post they liked as the likes arrive.

No worries, I was curious to see what code consuming the firehose looked like, both in terms of actual connection and how the auth works

If you have any links to example projects or whatever that would be amazing, I have some ideas for a firehose