chamuy0 von tsu

chamuy0 reads your LinkedIn feed and puts a small chip on every post: the slop score out of 10 and the verdict. Click the chip to see the signals behind the score — engagement bait, humblebrag, AI generic, buzzwords, broetry, fabricated story, manufactured hook, rage bait.

You can train it: label a post clean, borderline or slop and the next classifications use your labels as reference. Everything stays in your browser.

Feed cleanup: set a threshold and posts at or above it get banished with a stamp animation into a thin marker with a Show button. Nothing is deleted, everything is reversible.

Requirements: a TypeSafe API key for Jev, the decision model that does the scoring. Get one at typesafe.ai. Cost is about $0.042 per million input tokens; a full day of scrolling costs a fraction of a cent.

Privacy: your key and labels never leave your browser profile. The only thing sent out is the text of the post being scored, to api.typesafe.ai. No analytics, no accounts, no server.

Open source, MIT licensed: https://github.com/tsu-ld/chamuy0

Currently works on LinkedIn only. More platforms planned.