Question 1

What does Jev Lens do?

Accepted Answer

It badges every X and LinkedIn post READ / MAYBE / SKIP as it scrolls into view, decided by Jev against an interest profile you write in plain English. READ posts get a green border and SKIP posts fade out. The README calls it your feed, judged in 300 ms.

Question 2

Why not use keyword filters or ask an LLM about every post?

Accepted Answer

The README argues keyword filters cannot tell a build log from a certificate announcement, and asking an LLM about every post is slow and expensive. Jev returns typed decisions with calibrated probabilities in about 300 ms at $0.042 per million input tokens with output free, which the README prices at roughly 3 cents per 1,000 posts.

Question 3

How do I install it?

Accepted Answer

It is not on the store. Download or clone the repo, open chrome://extensions, turn on Developer mode, click Load unpacked and pick the extension/ folder. Chrome will not install extensions from outside the Web Store as .crx files, so Load unpacked is the intended path.

Question 4

How do I configure what I want to see?

Accepted Answer

Open the Jev Lens popup: it is the whole configuration screen. Write who you are in one sentence, posts you want to read with topics and example posts, posts you want to skip, and thresholds for how sure Jev must be. Paste your TypeSafe API key, click Test, then Save and reload the feed tab.

Question 5

What do I need before it works?

Accepted Answer

A TypeSafe API key: paste it in the popup, click Test — it should say key ok · models: jev-latest, jev-preview — then Save. Until a key is saved, posts show a small red 'jev: no API key' badge instead of a verdict. Jev was in early access at the time of writing, so a key may need the waitlist.

Question 6

Does anything leave my browser?

Accepted Answer

The README's pitch is your API key, your profile, your browser, with no scripted rules, no keyword lists and no cloud account of the project's. Judgments go to Jev with your key; ads and promoted posts are skipped without a call. The README does not describe any other data collection.

Question 7

Is there a deeper explanation of how Jev and this extension work?

Accepted Answer

Yes. The README links a companion blog post by the author, Jev and RLCD: The Model That Never Speaks, and Why Testers Should Care, which covers how Jev and RLCD work, why System 1 models fit automation better than chatbots, and what he learned building the extension.

Jev Lens

Highlights

Questions about this entry

Reactions & coverage

From the Jev guides

From the Jev AI model community