Agent routing & tool choice — page 4

Not every coding task needs your best model. Experimental Jev-powered model routing for Claude Code — V3 prototype runs today, V4 routes at the task boundary.

Shift every LLM call to the cheapest model that can handle it. Routing decided by TypeSafe Jev in ~180 ms. No training data. Policy in plain YAML. TypeScript and Python.

Route models and tools with TypeSafe

Task coordination for Codex CLI and Claude Code: carry bounded work directly or dispatch Herdr-backed sessions rooted in each app, from single-app repos to monorepos.

Benchmarks of MCP tool selection accuracy and latency. V2 evaluates 143 Woku tools with GPT-6 Luna API calls and Jev; includes Go/Rust cores, one TypeScript SDK and reproducible reports.

Keep skill catalogs outside the main LLM context. Jev selects relevant skills through one read-only MCP tool.

spaCy vs TypeSafe Jev on English spam, support routing and Swedish routing, evaluated in Braintrust

pi-thinking-router-jev —— pi 的自适应 Thinking Level Router。在任务起点、推理型失败与稳定窗口咨询 jev 决策模型（未配置时本地 规则兜底），动态选择 low/medium/high/xhigh 推理档位：简单改动降档、复杂 debug 升级、失败可升级、稳定可降级。含防抖钳制 、失败安全回退与 JSONL 决策日志，不切换模型、零侵入 agent loop。

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.

Jev-powered model and reasoning-effort router for Codex in DeepSeek Harness

Message triage CLI in Go, built on the Jev decision model from TypeSafe AI. Categorizes messages, scores urgency, and flags low-confidence ones for human review.

DSH plugin: register TypeSafe Jev (System One decision model) as an agent tool — jev_decide returns calibrated probabilities (noul/choice/score) for routing/triage/guardrail judgments, no text generation. 把 TypeSafe Jev 决策模型注册为 DSH agent 工具

Per-turn router for agent harnesses: one 350ms Jev call picks the model tier, effort, tools and skill, behind a hard deadline with a regex fallback. Claude Agent SDK adapter included.

Tell Claude Code which installed skill a session needs, using Jev (TypeSafe AI) for the decision and skills.sh for discovery.

Benchmarking TypeSafe's Jev decision model as a cost-efficient LLM router on RouterArena

A .NET 10 and React 19 application for fast, structured AI-powered ticket triage using TypeSafe Jev.

Type-safe model router. Jev (System One) banks each request to a typed catalog route.

Tag $agent inside your Pull Requests and Issues to let Jev decide which model and reasoning to use.

Experimental native Codex Desktop and CLI model routing with Jev and a configurable allowlist

A fast System One decision model (Jev or local Laya) picks which LLM answers each prompt: OpenAI-compatible Go gateway + decision benchmark

Fail-open Codex UserPromptSubmit hook that injects TypeSafe Jev pre-task routing metadata.

Portable Codex Skill for Jev model and reasoning-effort routing, with safe installation and Chinese usage guides

钉钉信息 Jev 分拣器（jev-dingtalk）：把钉钉邮件与聊天分拣成 Now / Today / Queue / Ignore 清单——谁在等回复、谁需要人看一眼。dws 取数、Jev 分类、报告本地渲染，可直接作为 Agent 技能使用。| DingTalk mail & chat triage with Jev.

Executable cognitive architecture: validated graph context, deterministic cache, bounded JEV routing and evidence-gated model collaboration.