yoshi

Context-pruning proxy for Claude Code and Codex: Jev judges which history is still needed, measured not claimed. POC here now, heading soon into https://github.com/compozy/compozy

How it uses Jev

Jev judges new context once, above a configurable size gate. Yoshi applies the validated omissions while keeping the protocol intact, then forwards the request to Anthropic or OpenAI and streams the response back unchanged.

Highlights

  • Judges new spans once : completed Bash , Write and Edit arguments, tool results and teammate reports. The default starts above 50,000 estimated input tokens. Previously forwarded…
  • Asks Jev, one span per request , with up to eight requests in flight. Each
  • Applies only validated omissions. A removed span becomes an explicit
  • Serves both protocols on one loopback listener: the Anthropic Messages

Source: community catalog. Details and metrics reflect the published community snapshot and are not a performance endorsement. Documentation details extracted from GitHub · the README.

Keep exploring

Similar to yoshi

Browse all
pi-jev-context project preview
Agent tooling
Reversible context pruning for Pi, powered by TypeSafe Jev. Keep useful context without deleting session history.
jev-skills project preview
Agent tooling
Skills without the context tax. Claude Code and Codex plugin: TypeSafe's Jev decides on every turn which skills the model sees. Always-on context cost: 0 tokens.
Jev Auto Router project preview
Agent tooling
Per-call GPT model routing for Codex: Jev chooses model and effort, a local Responses proxy keeps the tool loop intact, then independent verification and Router Compass check whether the task still passed.
jev-shell-history project preview
Integrations
Fish-style zsh history autosuggestions ranked by Jev (TypeSafe)
got-jev project preview
Demos & playgrounds
Jev (TypeSafe AI) PoC through Game of Thrones
jev-playwright-mcp project preview
Browser & computer use
Jev-augmented Playwright MCP proxy — page-state triage, prompt-injection shielding, goal-based snapshot pruning, risky-action gating. Drop-in wrapper around @playwright/mcp for any coding agent.
eve project preview
Integrations
Vercel's open agent framework, which ships Jev as the default evaluation model in its experimental evaluate path.
barrunto project preview
Browser & computer use
A Chrome extension that brings TypeSafe's Jev to X.com to analyze posts as you browse
agent-fastpath project preview
Browser & computer use
Jev MCP server: a decision layer for coding agents, built on TypeSafe Jev (System One model). Ship gates, risk checks, file triage that keeps files out of context, and a safe headless browser, with calibrated confidence. For Claude Code, Codex, Cursor.