User repository on GitHub
Map a codebase into units and let Jev (TypeSafe AI) hand an AI coding agent the ten files that matter for a task
- Forks
- 1
- Last pushed
- Updated Sep 2026
Repo stats from the GitHub API, cached Oct 2026.
CommunityJev Agents & MCP
Jevmap
Codebase mapper that indexes a repo into small units without AI, then asks Jev which units matter for a plain-language task and returns ranked files with line numbers and probabilities.
- Stars
- 1
- Language
- JavaScript
- License
- MIT
- Published by
- Community
- Author
- aegsrl7
- Last verified
- Oct 2026
About this project
jevmap builds a map of a codebase as small units (functions, endpoints, classes, components, pages, jobs) by reading the source, without AI, and then asks Jev (TypeSafe AI) which units matter for a task written in plain language. The result is a short list of files, with line numbers and probabilities, that an AI coding agent reads before touching anything. Zero runtime dependencies, Node.js 18 or newer, MIT.
Quoted from the project's README. Read the full README.
Works forJev for Tool calling
Used inJev for Software & DevTools
- Languages
- JavaScript
Highlights
- Build step needs no key or network, reading tracked files or honoring .gitignore; only find and bench call the Jev API.
- Find prints the task, scan stats, files to read, and top units with probabilities; JSON and files-only output suit agent consumption.
- Prefilter mode narrows candidates first for faster cheaper searches; a bench command measures recall against recent commit history.
- Zero runtime dependencies on Node 18+, with optional Anthropic assistance for splitting tasks and filling unit descriptions.
Watch out
Ranking quality depends on the repo and unit descriptions; measure your own project. 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.
No. The README stresses numbers vary by repository and description quality, and ships a bench command that replays your own commits as tasks so you measure your repo directly.
Scan judges every unit for the best recall signal, while prefilter narrows candidates first for faster cheaper runs; bench both on your history and pick from your own table.
Building the map still works offline, but find falls back to printing the keyword prefilter order with scores instead of probabilities and says so.
Your choice: gitignore it for a local-only index or commit it so every agent and teammate shares the same map.
It scans the repository and writes map.json plus a readable map.md index, one section per area and one bullet per unit. Files come from git ls-files when git is available, otherwise a directory walk that honours .gitignore, and building the map needs no network or key.
build, find, describe, bench, and help/version. find supports scan or prefilter mode with a top count; describe fills missing unit descriptions with an Anthropic model; bench replays recent commits as tasks to measure your own repository.
MIT, with zero runtime dependencies and Node.js 18 or newer. jevmap.config.json is optional and every field in it is optional, with command-line options overriding it; exit codes are 0 ok, 1 error, and 2 usage.
More Jev resources like this
3GitHub stars
MCP server that exposes Jev classify, score, yes-no check, risk gate, and multi-question decide tools to coding agents through one Jev call per tool.
Jev Agents & MCP#community#github
Community
codaaiteam/jev-mcpJavaScript
codaaiteam/jev-mcpJavaScript0GitHub stars
Demo pairing Vonage Video API with Jev: chat sentiment and frustration scoring offers a video call and pings an employee dashboard.
Jev Agents & MCP#community#github
0GitHub stars
Agent plugin adding typed Jev judgment packs to six existing coding skills for Codex, Claude Code, and Oh My Pi, without changing their procedures.
Jev Agents & MCP#community#github
118GitHub stars
A Claude Code skill that moves bulk judgment off Claude and onto Jev. You filter files, classify items, rank candidates or locate lines, and it returns typed verdicts, a shortlist with borderline markers, and a receipt of the tokens Claude did not read.
Jev Agents & MCP#community#jevbest
Community
UditAkhourii/quicksilverJavaScript
UditAkhourii/quicksilverJavaScript47GitHub stars
MCP server that screens files, public webpages, or inline text against a query and returns compact relevance probabilities, so the main agent opens only promising items.
Jev Agents & MCP#community#jevbest#catalog-import
Community
kbhuw/jev-siftJavaScript
kbhuw/jev-siftJavaScript16GitHub stars
Standalone terminal agent that puts Jev inside its planning loop, proposing text-model steps and scoring candidates against independent rubrics before a deterministic policy executes one.
Jev Agents & MCP#community#jevbest#catalog-import
From the Jev community
Posts from builders shipping with Jev right now.