jevgrep (dzhng)
Code search CLI and agent skill (jg) where Jev judges relevance across folders, files and declarations to return excerpts.
Overview
jevgrep (dzhng) helps coding agents start unfamiliar tasks by finding code from a description of what it does. Ask jg a repository question and it returns relevant files, reading leads and verbatim source excerpts in one stdout response, using Jev to judge relevance hierarchically across folders, files and declarations. The calling agent then implements and tests the change. The README's headline is same intelligence at about 30% lower cost: in the author's ten-task SWE-bench comparison, agents with jevgrep completed the same 8 of 10 tasks as the baseline at lower cost.
It requires Node.js 22+ on macOS, Linux or Windows and a key for Vercel AI Gateway, TypeSafe, OpenRouter, OpenCode Zen or a custom TypeSafe-compatible endpoint, with provider selection from version 0.3.0. No Python, Bun or ripgrep install is needed. A separate skill must be installed for agents to know when and how to use it; jg skill detects Claude Code, Codex, OpenCode and other agents and installs it, and the skill checks for and installs the CLI if missing. It is unrelated to other projects named JevGrep, is published on npm and is MIT licensed.
What the decision model decides
| Type | Question | Answers |
|---|---|---|
| Score | How relevant is this folder, file or declaration to the question?(inferred) | — |
How it works
- Install CLI and run jg auth and jg skill
- Agent asks jg a repository question
- Jev judges relevance across folders, files and declarations
- Return files, reading leads and verbatim excerpts
Strengths
- README reports same task success at about 30% lower cost
- Skill teaches agents when to use it
- No extra runtime dependencies
Caveats
- Ten-task comparison is small
- Provider key required
Getting started
npm install -g @dzhng/jevgrep, then `jg auth`, `jg skill`, and `jg "How are telemetry events recorded and sent?" ./my-project`.At a glance
- Type
- Agent skill
- Status
- beta
- Language
- TypeScript
- Runtime
- Node CLI / agent skill
- Has code
- yes
- License
- MIT
- Stars
- 2,671
Use cases and integrations
How curated lists describe it
Code search CLI and skill for coding agents: Jev judges relevance across folders, files, and declarations, and `jg` returns files, reading leads, and verbatim excerpts. In the author's ten-task SWE-bench comparison it solved the same 8 tasks as the baseline at about 30% lower cost.
AnotiaWang/awesome-decision-modelsFind code by asking what it does. A CLI for coding agents that uses Jev to discover relevant files and source context with exact excerpts and line numbers.
logicrw/awesome-jev-projectsFind code by asking what it does. A CLI for coding agents that uses Jev to discover relevant files and source context.
wh000wh000/awesome-jev-liveFind code by asking what it does. A CLI for coding agents that uses Jev to discover relevant files and source context.
hellogumbo/awesome-jevFinds code by what it does rather than what it matches: a CLI for coding agents that asks which files and which regions are relevant to a description. Unrelated to the JevGrep above, which has the same name.
MrJev/awesome-jev
Frequently asked questions
Does it save money?
In the author's ten-task SWE-bench test it matched baseline success at about 30% lower cost.
Do I need the skill?
Yes, installing the CLI alone does not teach your agent to use it.
Which providers work?
Vercel AI Gateway, TypeSafe, OpenRouter, OpenCode Zen or a custom compatible endpoint.
Related projects
Local code search for coding agents: ripgrep/BM25 shortlist, then Jev judges each chunk's relevance, served over MCP.
MCP serverbetaRustCLI and MCP server giving coding agents Jev-powered semantic search, focused reads and list picking in about a second for about a cent.
BenchmarkbetaPythonSkill and MCP service that audits an app for AI calls that are really decisions and shadow-tests migrating them to Jev.
MCP serverbetaTypeScriptMCP server that returns only the JS/TS code a task needs by following the TypeScript compiler graph and asking Jev about relevance.
MCP serverbetaTypeScriptMCP server giving coding agents fast typed decisions (ship gates, risk checks, file triage, page checks) as accept, review or escalate.
MCP serverbetaTypeScriptSingle MCP tool that indexes a repo and has Jev rank a shortlist so a coding agent reads the right files and lines instead of grepping.
MCP serverbetaPython