Skip to content
JevDirectory.orgTraffic count unavailable. View site traffic
CommunityJev Practices & Patterns

Jevalyzer

Grades the coding agent sessions already sitting on your disk. Jevalyzer reads Claude Code, Codex CLI, opencode, Gemini CLI and Antigravity transcripts, scores each exchange, and renders an interactive HTML report or terminal dashboard.

See highlightsOpen github.com
Stars
3
Language
TypeScript
License
MIT
Published by
Community
Author
killerz3
Last verified
Oct 2026

About this project

Product page · AGENTS.md (let an agent install it) Grade the agent sessions already sitting on your disk. Jevalyzer reads the chat logs your coding agents write locally — Claude Code, Codex CLI, opencode, Gemini CLI and Google Antigravity — scores every exchange, finds where the agent went wrong, works out which model you actually prefer, and renders it as an interactive HTML report or a terminal dashboard.

Quoted from the project's README. Read the full README.

Tagscommunityawesomejevcatalog-import
Works forJev for ClassificationJev for RankingJev for Tool calling
Used inJev for Software & DevTools
Languages
TypeScript

Highlights

  • bun start finds your sessions, shows the token and dollar estimate before sending anything, then scores with a live progress display.
  • Two profiles share one store: minimal at 7 questions, extensive at 21, and an extensive row satisfies a later minimal request.
  • Results are cached by content hash, so re-running pays only for exchanges it has not seen and Ctrl-C is always safe.
  • Backends are Vercel AI Gateway, Cloudflare Workers AI and TypeSafe directly, and --model accepts any AI SDK evaluation model.

Watch out

Requires Bun 1.2 or newer; the session readers use bun:sqlite, so Node alone will not do. Not on npm yet, so the clone is the install path. Your transcript text leaves the machine: prompts, replies, tool calls and tool output go to the gateway for scoring, with nothing else uploaded, nothing stored outside ~/.jevalyzer and no telemetry. Backend defaults have real catches. On a hobby Vercel plan a card must be on file before it serves anything, free-tier requests on this model are throttled well below the documented limit so runs slow rather than fail, and zero data retention needs Pro or Enterprise so it is dropped automatically after the first rejection. Cloudflare serves Jev as a partner model outside the free Neurons allocation, so it returns an insufficient-balance 2021 and is never auto-selected, and its context is 32k rather than 64k. The cost table is measured against one 40 MB Claude Code archive, with results cached by content hash and a default $5 budget ceiling that requires confirmation to exceed. 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

tenbin
4GitHub stars
Design-time toolkit for using Jev from a coding agent, combining docs, a stdio MCP server, and an agent skill for decomposing judgments into Choice, Score, and Noul.
Jev Practices & Patterns#community#awesomejev#catalog-import
Communitysimota/tenbinTypeScript
jev-x-kit
3GitHub stars
A TypeScript framework that puts a System 1 decision layer in front of agent micro-decisions, shipped as a Claude Code plugin, an MCP server and a CLI, with an offline simulator.
Jev Practices & Patterns#community#awesomejev#catalog-import
CommunityKadihx/jev-x-kitTypeScript
jev-carryforward
2GitHub stars
An MCP server with two tools for agent memory: record writes facts into a per-project ledger as they happen, and recall scores each one against the task you are starting and hands back only what matters.
Jev Practices & Patterns#community#awesomejev#catalog-import
CommunityDharundp6/jev-carryforwardTypeScript
Kev
20GitHub stars
Self-hosted Apache-2.0 decision engine with a Jev-style wire format that turns a model you point at it into typed choice, score, and noul answers.
Jev Practices & Patterns#community#awesomejev#catalog-import
Communityarjun988/KevTypeScript
claude-jev
7GitHub stars
Claude Code plugin that routes code review, debugging, planning, and search triage through typed Jev judgements with fixed thresholds applied in code.
Jev Practices & Patterns#community#awesomejev#catalog-import
Communitybuchmark/claude-jevTypeScript
jev-agent-authorization
5GitHub stars
Starter kit putting Kinde identity plus Jev judgment in front of every agent tool call, with allow, hold-for-approval, or deny decided in code and written to a ledger first.
Jev Practices & Patterns#community#awesomejev#catalog-import
Communitykinde-starter-kits/jev-agent-authorizationTypeScript

From the Jev guides

Original write-ups on what this entry is for.

Designing the State: What to Include and What to Drop
Most Jev accuracy problems are state problems. What belongs in a state, what is a pre-made judgment to leave out, and why naming the slice you ask about matters more than the wording.
Jev Practices & Patterns#state#design#primitives
Read article
Noul, End to End: One Probability and a Dozen Thresholds
A Noul returns one number and no separate confidence. How to read the value, pin its boundary with criteria, split compound questions, and set a per-action threshold that matches the cost of being wrong.
Jev Practices & Patterns#noul#primitives#confidence
Read article
Scoring Without Measuring: Bands, Rubrics, and Composite Triage
A Score is a weighted position, not a ruler. Why the number can land between levels, why levels are judged separately, and how to combine several atomic scores in code instead of one crowded rubric.
Jev Practices & Patterns#score#primitives#scoring
Read article
Back to all resources

From the Jev community

Posts from builders shipping with Jev right now.

Follow @typesafeai