Skip to content
Metatext Decide — typed decisions for your agents, with a probability on every answerGet early access
Metatext
Sign Up

langgraph-jev-router

LangGraph decorator that lets Jev take routing decisions it is confident about and falls back to your existing LLM router otherwise.

Librarybeta
View on GitHub ↗

Overview

langgraph-jev-router puts Jev in front of routing you already have in LangGraph. Adding the @jev_first decorator to an existing conditional-edge router means that when Jev is confident it decides and your LLM router is never called; when Jev is unsure, unavailable or mid-conversation, your router runs exactly as before. The existing router becomes the safety net, so adoption cannot make routing worse and nothing breaks if Jev fails.

An optional attach_routes call lets Jev read node docstrings as route descriptions for better accuracy. It supports conditional-edge functions, nodes returning Command(goto=...), and code you cannot edit such as prebuilt supervisors or compiled subgraphs via an install_shortcut helper. A recommended rollout runs Jev in shadow mode before going live. Without a key, examples use a deterministic keyword-based FakeJev and say so, and 59 tests run offline; a check command verifies key, network and model.

What the decision model decides

TypeQuestionAnswers
ChoiceWhich route should this message take (e.g. billing, tech, account)?billing · tech · account

How it works

  1. Decorate the existing router with @jev_first
  2. New message triggers a Jev routing question over the route names
  3. If confident, Jev's route is used
  4. If unsure, unavailable or mid-conversation, the original router runs
  5. Optionally run in shadow mode first

Strengths

  • Cannot make routing worse than the existing router
  • One-line adoption
  • Shadow-then-live rollout
  • Offline tests and fake client

Caveats

  • Benefit depends on how often Jev is confident
  • Requires a TypeSafe key for real use
  • LangGraph-specific

Getting started

Decorate your router with @jev_first, export TYPESAFE_API_KEY, and run jev-router check.

At a glance

Type
Library
Status
beta
Language
Python
Runtime
LangGraph library
Has code
yes
Has tests
yes

Use cases and integrations

How curated lists describe it

Frequently asked questions

What if Jev is wrong or down?

Unsure, unavailable or mid-conversation cases fall through to your existing router.

Do I have to change my router?

No, you add one decorator; your router stays as the fallback.

Can I test without a key?

Yes, examples use a deterministic FakeJev and tests run offline.

Related projects

  • tiershift

    Policy-driven router that sends each LLM request to the cheapest capable model tier, escalating when Jev finds a reason.

    ProjectbetaTypeScript, Python
  • AgentScope

    Alibaba's AgentScope 2.0 agent framework, which added a JevClassifierModel for binary, choice and score decisions via the TypeSafe SDK.

    ProjectproductionPython
  • HiRoute

    Local-first routing engine for long-running agent tasks that uses a Jev decider to pick a model group per turn.

    PluginproductionRust, TypeScript
  • LiteLLM

    Open-source AI gateway for 100+ LLMs that can use Jev to classify requests for complexity routing and to compact tool results.

    ProxyproductionPython
  • openchamber

    Open-source workspace for running and reviewing AI coding agents across devices, with an optional Jev-based automatic model router.

    ProjectproductionTypeScript
  • jev-router★ 593

    Per-turn model routing for Claude Code and Codex: wraps the real CLI and lets Jev send simple turns to fast models and hard ones to strong.

    ProxybetaJavaScript