Skip to content
JevDirectory.org
CommunityTools & Integrations22 starsVerified 2026-09-22

Yoshi

An experimental local context-pruning proxy for Claude Code and Codex that asks Jev, through the Vercel AI Gateway, which history spans to omit, then forwards a rewritten request with omission markers.

Open github.comcompozy/yoshi
Category
Tools & Integrations
Published by
Community
Author
compozy
Added
2026-09-22
Tagscommunitytypescriptcompactioncoding-agentclaude-codecodex

Highlights

  • Judges completed Bash, Write, and Edit arguments, tool results, and teammate reports once, above a default 50,000 estimated-token gate.
  • Each span is one Jev request, up to eight in flight, asking whether a required fact or a binding user constraint would be lost.
  • Both loss probabilities must be at most 0.2 to omit an input or report span; tool results keep the Atomic relevance/coverage/conflict rule.
  • Serves the Anthropic Messages API and the OpenAI Responses API on one loopback listener at 127.0.0.1:8790.
  • The README reports both v12 trials were slower than baseline and that a Jev outage can hold a request up to the 45-second deadline.

Quickstart

bash
bun install
cp .env.example .env   # new installations only; set the Gateway key and policy knobs

Watch out

MIT-licensed, but an experimental proof of concept whose successor is CompozyOS; needs Bun and a Vercel AI Gateway key, and candidate spans including source code go to TypeSafe with no zero-data-retention option.

More like this

save-token-jev
63GitHub stars
A Jev-guided context compaction library for coding agents: instead of summarizing old context, Jev decides which tool calls and results still matter while user and assistant text is kept verbatim.
Tools & Integrations#community#typescript#compaction
Communitysave-token-jev-clean
Fast Jev Compaction
6.2kGitHub stars
A Claude Code plugin and npm library that replaces transcript summarization with Jev keep-or-drop judgments, deleting stale tool calls and results while kept text stays verbatim.
Tools & Integrations#community#typescript#claude-code
Communityfast-jev-compaction
Jev Review
201GitHub stars
A local-first MCP server that gives Claude Code, Codex, Cursor, and OpenCode structured quality scores from Jev across correctness, complexity, tests, security, and other dimensions while the agent writes code.
Tools & Integrations#community#typescript#mcp
Communityjev-review
Back to all resources

From the community

Posts from builders shipping with Jev right now.

Follow @typesafeai

A really smart switch statement

Nathan Flurry 🔩
Nathan Flurry 🔩
Rivet
@NathanFlurry
·

hype-free explanation of jev: jev does not replace gpt / claude jev is just a *really* smart switch statement like if 2016 ml classifiers got 2026 levels of intelligence it's a new* type of tool that will make a lot of workloads insanely fast, cheap, and accurate * = and by Show more

Diogo Almeida
Diogo Almeida
TypeSafe AI
@CompleteSkeptic

After co-inventing ChatGPT, I kept asking myself: why have superhuman chat models not led to AGI? I’ve spent the last 2 years in stealth building a new way to train models (RLCD), and a new type of frontier AI model that we are releasing today: Jev • 20-200x faster • 40-400x

Watch on X
Reply
Read 230 replies

When a designer gets Jev

Full Jev video tutorial

The case against Jev-scored compaction

Theo - t3.gg
Theo - t3.gg
@theo
·

This is a terrible compaction strategy that fundamentally doesn't understand how compaction and context management work. Seems like a lot of people are confused so let's break this down. 1. Compaction isn't a filter The role of compaction is to clean up history to keep the Show more

tamara
tamara
@tamarajtran

found the perfect use case for @typesafeai Jev: instant compaction in 2026, why is compaction still a summarization prompt? Jev can make it instant by scoring every tool call and dropping what’s irrelevant

Watch on X
Reply
Read 182 replies

Classifying rows in DuckDB

A playable 16-judgment demo