vibeCheck

a browser extension that filters out bad vibes from ur social media feed. it scans posts in real time using nlp and if something is negative/toxic/angry it blurs it out so u dont have to see it unless u want to.

EXTENSION TESTED VIDEO: https://drive.google.com/file/d/1XTuXyc3mkDoZaKwPaStXoCQ5Hw5V525x/view?usp=sharing

Time lapses:

https://lapse.hackclub.com/timelapse/_X-6potlNUpJ https://lapse.hackclub.com/timelapse/bY8CkBlyZ6nz https://lapse.hackclub.com/timelapse/FKQTmGNMvgNA https://lapse.hackclub.com/timelapse/1irqj6JTP7Rh https://lapse.hackclub.com/timelapse/D0z9hsD2PPFp https://lapse.hackclub.com/timelapse/9XnyvLFREmfY https://lapse.hackclub.com/timelapse/ded5YrtHP_Le https://lapse.hackclub.com/timelapse/5D6CK0iHUHvr https://lapse.hackclub.com/timelapse/DYcb57hDh1lZ

what it does

• real time scanning - watches ur feed as u scroll, picks up new posts automatically using mutation observer on the dom
• sentiment analysis - runs every post thru a scoring engine with weighted word dictionaries, negation handling, all caps detection etc
• blur overlay - if a post scores below ur threshold it gets blurred with a warning like "potentially negative content: Anger" and a reveal button if u still wanna see it
• category toggles - choose what u wanna block: anger, sadness, toxic. mix and match however u want
• whitelist - add usernames or keywords that should never get blocked, even if the content is negative
• stats dashboard - see how many posts got filtered and what categories they were in
• works on twitter/x, reddit, and linkedin - each platform has its own dom selectors with fallbacks

setup

1. clone this or download it
2. go to chrome://extensions
3. turn on developer mode
4. click load unpacked and select this folder
5. thats it, go to twitter/reddit/linkedin and it should be working

the popup

click the extension icon to open the settings dashboard where u can:

• enable/disable the whole thing
• adjust sensitivity with the threshold slider (-5 to 0, default -2.0)
• toggle which emotions to block
• add whitelist entries (users or keywords)
• see ur filter stats
• clear stats whenever

supported sites

site	selector	backup selector
twitter/x	[data-testid="tweet"]	-
reddit	shreddit-post	div[id^="t3_"]
linkedin	div.feed-shared-update-v2	div[data-id^="urn:li:activity"]

how the scoring works

each word gets looked up in the dictionary. bad words have negative scores (-0.6 to -1.8) and a category (anger/sadness/toxic). good words have positive scores (0.7 to 1.5) and offset the bad ones.

if the previous word is a negation word (not, never, no, hardly, etc) the score gets flipped by -0.5x.

if the whole post is ALL CAPS and has bad words the score gets a 1.5x multiplier.

final score gets compared to ur threshold. if its below and matches a blocked category = blurred.

hope yall like it lol!