LinkScout
Browser extension that has Jev score search results for relevance, depth, SEO filler and page type, then re-ranks them with a 0-100 badge.
Overview
LinkScout scores the links on a search results page before you click them. An MV3 browser extension watches Google, Bing and DuckDuckGo results, including ones that load later, and sends the query and URLs to a Cloudflare Worker. The worker fetches each page as Markdown with Cloudflare Browser Rendering, falling back to Jina Reader, and asks Jev, in one request per result, for relevance and depth scores, yes/no checks for SEO filler and sales pages, a page-type choice, and a choice over pre-split paragraphs to find the key passage. Code combines these into a 0 to 100 badge and re-ranks the results.
Results are cached for 24 hours per query and URL. The key passage is shown alongside the badge, with on-device translation into Traditional Chinese when that interface language is used. Deployment needs Cloudflare credentials, a TypeSafe key and optionally a Jina key, and the README warns that the production token must be random because the worker URL is public and otherwise strangers could spend your quotas.
What the decision model decides
| Type | Question | Answers |
|---|---|---|
| Score | How relevant is this page to the query? / How deep is its content? | — |
| Yes / no | Is this page SEO filler? / Is it a sales page? | — |
| Choice | What type of page is this? | — |
| Choice | Which paragraph is the key passage for the query? | — |
How it works
- Extension collects result URLs and the query
- Worker fetches each page as Markdown
- Jev answers relevance, depth, spam, type and key-passage questions
- Code combines them into a 0-100 score
- Extension shows badges and re-ranks results
Strengths
- Judges actual page content, not snippets
- Key passage surfaced before clicking
- 24-hour caching limits cost
Caveats
- Requires deploying a Cloudflare Worker with several keys
- Public worker URL must be protected with a random token
- Page fetching can fail on some sites
Getting started
npm install, put keys in .env, npm run dev for the worker, then load extension/ unpacked in Chrome.At a glance
- Type
- Browser extension
- Status
- prototype
- Language
- JavaScript
- Runtime
- MV3 browser extension and Cloudflare Worker
- Has code
- yes
Use cases and integrations
How curated lists describe it
Web search: browser extension that asks Jev, in one request per result, for relevance and depth `Score`s, `Noul`s on SEO filler and sales pages, a page-type `Choice` and a `Choice` over pre-split paragraphs for the key passage, then combines them in code into a 0–100 badge that re-ranks Google, Bing and DuckDuckGo results.
yibie/awesome-jev
Frequently asked questions
What does Jev decide?
Relevance, depth, whether a page is SEO filler or sales, its type, and its key passage.
Which search engines are supported?
Google, Bing and DuckDuckGo.
What do I need to deploy it?
Cloudflare credentials with Browser Rendering, a TypeSafe key and optionally a Jina key.
Related projects
- Jev Search★ 534
Plain-language web search where a decision model like Jev picks query, sources and time range, then ranks Search1API results.
AppproductionTypeScript Chrome extension that re-ranks Google results with Jev questions per result and folds away sales pages and SEO filler.
Browser extensionbetaTypeScriptResearch agent where GPT plans and Jev finds evidence in parallel across document regions with Choice and Noul.
Browser extensionresearchTypeScriptshadcn/ui site-search block that shows keyword hits instantly, then has Jev re-rank the top 20 by what the visitor meant.
DemobetaTypeScriptWeb search CLI for agents that has Jev score results and scraped chunks against the query and goal, passing only relevant ones on.
BenchmarkbetaGoMilvus bootcamp notebooks combining Gemini embeddings, Milvus retrieval and Jev judgments for reranking, filtering, routing and stopping.
ArticleproductionPython