Does it crawl or scroll my feed on its own?

No. It reads only the tweets already loaded in the open X page, sends their text plus your goal and judgment subject for a structured judgment, and puts an icon, percentage and color on each. It does not search, scroll or open posts; clicking a recommended candidate jumps you to it, and the top-3 comparison never reorders the page.

What exactly gets sent to the API, and what does it cost?

The tweet text, your goal and the judgment subject go to the TypeSafe API, and your key is kept in extension local storage. The README says not to use it on content you have no right to transmit, and that pricing and quotas are the provider's to define. The optional top-3 mode costs extra tokens on top of per-tweet judgments.

Can I install this like a normal extension?

Not from a store. It is a Manifest V3 extension with no build step, loaded unpacked from chrome://extensions with developer mode enabled, picking whichever folder actually contains manifest.json. Publishing through the Chrome Web Store, and the maintenance cost of it, are still unchecked roadmap items.

Should I trust the percentages shown in the screenshots?

The README says not to. The images were produced for illustration rather than captured from an actual run, and the 86% figure displayed is not a measured model score. The project also frames the whole thing as auxiliary filtering rather than verification, and warns the probabilities are not real-world accuracy rates.

What does the optional best-match-three feature do?

When at least two candidates clear the threshold, it pairwise compares up to 12 loaded posts and recommends at most three with links. It does not reorder the X page, but it consumes extra API tokens.

Can I clear and later restore a screening setup?

Yes. Clearing stores the goal and audience in history and turns off auto-judging, and restoring fills the settings back and re-enables auto-judging, which may create new API usage.

What license is Tweet Radar under, and who runs it?

It is MIT-licensed and open source, not affiliated with X or any model provider, and the extension is currently side-loaded through Chrome developer mode rather than the Web Store.

CommunityJev Demos & Experiments

Tweet Radar

Free Chrome extension that scores tweets already loaded on your X page against criteria you write yourself, showing the probability as a badge. The substance is the prompts library of goals, judgment subjects and counter-examples; the extension is a no-build sideload with no store listing.

See highlights

Open github.com

Stars: 1
Language: JavaScript
License: MIT
Published by: Community
Author: kelaocai
Last verified: Oct 2026

kelaocai/tweet-radar

User repository on GitHub

View on GitHub

Jev / TypeSafe AI 驱动的 X 信息筛选：按你的规则发现值得读的帖子。开源判断规则库 + 免费 Chrome 扩展。Filter X with your own criteria.

Forks: 0
Last pushed: Updated Sep 2026

Repo stats from the GitHub API, cached Oct 2026.Homepage

Tagscommunity awesomejev catalog-import

Works forJev for Classification Jev for Moderation

Used inJev for Software & DevTools

Languages: JavaScript

Highlights

It reads only tweets the page has already loaded, and never searches, scrolls or opens posts on its own; clicking a result jumps to it.
The core is the prompts directory: copyable goals, judgment subjects and fictional counter-examples, plus worked examples of profiles.
Manifest V3 with no build step and no dev dependencies, sideloaded through Chrome developer mode rather than installed from a store.
An optional top-3 mode pairwise-compares up to 12 shortlisted candidates, costing extra tokens without reordering the X page.

Watch out

You need Chrome or Chromium and your own TypeSafe API key: the extension is free but the judgment service bills you, and the README defers plan, quota and pricing to the provider's own pages. Thresholds, batch size and the per-page cap are adjustable in settings, so the defaults are not fixed behavior. Tweet text, your goal and the judgment subject are all sent to the TypeSafe API, and the key lives in extension local storage, so do not run it over content you have no right to transmit. The project is explicit that this is a filtering aid, not fact-checking, investment advice or a guarantee of authenticity, and that the percentages are not real-world accuracy rates. Installation is sideloading, so developer mode and manual updates are on you, and store publishing is still an open roadmap item. The README's screenshots were produced as illustrations rather than captured from a run, and the 86% shown is not a real model score. Copy reviewed against the repository README on 2026-10-04, not from running it.

Questions about this entry

Short answers drawn from the linked project or page.

More Jev resources like this

last-exit

Last Exit is a short cyberpunk checkpoint game where you type a cover story, present a manifest and offer a bribe to an inspector who decides whether to let you through. Each reply makes two Jev calls: one turns your statements into typed observations, the other picks his next action.

Jev Demos & Experiments#community#jevbest#catalog-import

Community

0x963D/last-exitJavaScript

sloppy-jevs-extension

Sloppy Jev's is a Manifest V3 Chrome extension that blurs AI-generated prose and ads, classifying extracted page text through jev-latest on the TypeSafe API. Plain JavaScript and CSS, no build step, MIT licensed, loaded unpacked.

Jev Demos & Experiments#community#jevbest#catalog-import

Community

neddes/sloppy-jevs-extensionJavaScript

laya-vs-jev-arena

Side-by-side game harness where Laya running locally and Jev over API answer the same typed questions for snake, kombat, runner, flappy, tetris, and human-versus-model matches.

Jev Demos & Experiments#community#jevbest#catalog-import

Community

PromptEngineer48/laya-vs-jev-arenaJavaScript

jev-kit

Claude Code kit that guards PreToolUse calls with a free pre-filter plus one Jev question for ambiguous cases, and adds sub-agent sizing, rechecks, search, daemon speedups, and monitoring.

Jev Demos & Experiments#community#awesomejev#catalog-import

Community

jonathanavis96/jev-kitPython

jev-tetris

Live versus Tetris where Jev faces Claude Haiku 4.5, Gemini 3.8 Flash, or local open-weight Laya on the same seeded piece sequence; cleared lines arrive as garbage rows and the first to top out loses.

Jev Demos & Experiments#community#awesomejev#catalog-import

Community

trungdq88/jev-tetrisJavaScript

jev-chat-for-twitch

Chrome extension that adds a second Twitch chat column showing only messages Jev rates as matching your chosen intent, with every row linked to its probability distribution.

Jev Demos & Experiments#community#jevbest#catalog-import

Community

ethanplusai/jev-chat-for-twitchJavaScript

From the Jev guides

Original write-ups on what this entry is for.

Per-Tick Decisions in Games

Sep 28, 2026

A game tick is a state plus a legal action list. How Clash Royale, Minecraft, and the earlier Mario, Pokémon, StarCraft, and drone loops ask Jev one step at a time.

Jev Demos & Experiments#gaming#agents#control

Read article

Back to all resources

From the Jev community

Posts from builders shipping with Jev right now.

Follow @typesafeai