When and Where Actually to use Jev in your code base.
JevX Find where Jev fits in your codebase — and change it.
Your code is full of hardcoded rules that are really judgment calls: regexes that guess what an error means, keyword lists that guess what a customer wants, slice(0, 3) that guesses what's best. JevX finds them, scores each one three ways (your AI, TypeSafe/Jev, patterns learned from real Jev projects) and replaces the strong fits with a Jev decision (TypeSafe noul · choice · score) — keeping the old rule as the fallback, and proving it with your own tests.
Or from the AI you already use — Claude Code, Claude Desktop, Cursor, Windsurf, VS Code, Gemini CLI, Codex:
Claude Desktop users can also double-click jevx.mcpb (from the releases) and pick their project folder.
- Your key, your AI. JevX never ships a key; the terminal version uses your xAI / OpenRouter key. - Safety lives in the tool: only strong fits change by default, the old rule stays as the fallback, files you're editing are skipped, your tests run before and after, jevx undo restores. - Private by default. Code goes only to your own AI. --share (opt-in) sends generic patterns — never code, file or function names — so JevX learns which changes hold up.
Full docs: apps/jevx/README.md · website: site/ · changelog: apps/jevx/CHANGELOG.md
Folder What --- --- apps/jevx the jevx CLI + MCP server (published to npm) packages/engine read → assess → scorecard → edit → apply → share packages/gemini AI transports (xAI, OpenRouter) packages/analyzer, scanner, core local indexing, static candidates, secret scrubbing tests/, examples/jev-demo tests with mock AI + mo
From the project README.
Add the radar badge to your README — it shows your project was picked up by MCP Radar and links to this page:
[](https://mcp.liqiwa.com/s/vij-sameerb5--JevX.html)
Screens an AI agent's tool calls before they run and tool results before the agent reads them. An MCP proxy plus a hooks adapter for a client's built-in tools.
shitianfang/jev-useClaude Code / Codex / pi plugin that hands agent steps needing no text output to Jev (TypeSafe's judgment model) — measured p50 ~230 ms and ~$0.02 per 1,000 judgments, with typed escalation back to th
nexibeo/jev-browser-controlLet Claude code, chatgpt codex or control your own Chrome. Chrome extension + MCP server: Jev, TypeSafe's decision model, picks each click in ~0.5 s for a fraction of a cent. MIT, bring your own OpenR
PerryLink/jevcoreTypeSafe Jev for DeepSeek Harness, the Model Context Protocol, and plain Node: typed judgments instead of prose, offline by default.
lanlan0811/tianshu-mcp天枢 × AI-Agent 编排 MCP server：把 其他AI-Agent 当作执行工人，自动完成「开发 → 验收 → 失败返修 → 再验收」闭环。
rashedInt32/jev-mcpMCP server exposing TypeSafe Jev as typed, calibrated judgment tools: classify, score, check, batched ask. Ships as a Claude Code plugin.
The top new MCP servers of the week, every Monday. No spam, unsubscribe anytime.