Community · E2 · artifact verified
Keep YouTube learning-focused with typed judgments
FocusMaxx, a Chrome MV3 extension keeping YouTube learning-focused: tutorials, technical discussions and documentaries stay visible while unwanted entertainment is covered - driven by two plain-language fields and Jev typed judgments, keywords as hints only.
01 · Role in the system
What Jev does here
The interesting design honesty is what Jev does NOT do: keywords are optional interest hints, never literal gates, and clickbait titles are not filtered - the model judges what a video is, not how it markets itself. Two plain-language fields (what you want to see, what you want to hide) shape the typed questions, and Jev's calibrated answers decide cover-versus-keep per video. The repo is unusually documented for its size: architecture doc, demo runbook, test report, a rehearsal screenshot explicitly labeled as authored judgments rather than live results, and an install ZIP.
02 · Control boundary
Where Jev sits
Jev as the per-video keep-or-cover judge of a learning-focused YouTube: plain-language show/hide fields shape typed questions, calibrated answers drive the cover overlay, and keywords stay advisory so judgment is semantic, not literal.
Code owns the loop, permissions, thresholds, validation, and side effects. Jev owns only the bounded judgments described above.
03 · Known limits
What this evidence does not prove
- YouTube-specific; the rehearsal screenshot uses authored judgments, not live results.
- 0 stars; created 2026-10-10, single-maintainer.
- Show/hide field phrasing shapes judgment quality; no measured precision published.
04 · Attribution
Public sources
This is a Community record: the project was published by a third-party community author.
- krishnakant2607 ↗Community · github · public · checked 2026-10-11