Why stay in shadow mode for a week before going active?

Because the router is advice and the Auto-review rules are the boundary. Shadow mode plus record and review.py lets you compare Jev advice with bot behavior before anything depends on it.

Which file do I actually need to review before trusting it?

Review jev/questions.py first: it holds every question and every threshold. Router behavior changes there, not in the surrounding glue.

What overrides Jev when the model says proceed?

Code overrides the lane for account-external, critical-risk, irreversible, looping, over-attempt, or low-confidence calls, and done is checked first so finished tasks never re-run.

How do I set up the key without pasting it in chat?

Create a dedicated capped TypeSafe key, add the router bot, run jev-setup, place the key in ~/.typesafe/key via Agent Computer, then run smoke.py. Never paste the key in chat.

What are the four actions the router can return?

proceed (do it), dry_run (draft or simulate, then stop and show it), ask_human (stop and ask, with the reason), and skip (the task is done or pointless). bypass means the router is off and normal rules apply.

How big a state can I send to Jev?

About 32k tokens is the request budget, and the router truncates long fields rather than sending them whole, so pass summaries rather than full transcripts.

How do I point the specialist question at my own bots?

The specialist question names the ten bots from the Grok Bot Business Starterpack, so edit SPECIALISTS in jev/questions.py for your own roster; that same file holds every question and threshold.

CommunityJev Integrations & Adapters

Grok Jev Router

Dependency-free Python toolkit that puts Jev in front of Grok Bot, turning calibrated answers into proceed, dry_run, ask_human, or skip with hard-coded safety overrides.

See highlights

Open github.com

Stars: 1
Language: Python
License: MIT
Published by: Community
Author: colinmcdermott
Last verified: Oct 2026

colinmcdermott/grok-jev-router

User repository on GitHub

View on GitHub

Jev decides, Grok Bot executes, humans control irreversible actions. A decision router for Grok Bot built on TypeSafe AI's Jev.

Forks: 0
Last pushed: Updated Sep 2026

Repo stats from the GitHub API, cached Oct 2026.

About this project

Jev decides, Grok Bot executes, humans control irreversible actions. A small, dependency-free Python toolkit that lets a Grok Bot use TypeSafe AI's Jev as a decision layer: before an expensive or risky action, the bot asks Jev a fixed set of typed questions and code turns the calibrated answers into one of four actions. About a hundredth of a cent and a few hundred milliseconds per decision. Install as a Grok Bot template: Add Jev router to Grok Bot. Or build your own from grok-bot/.

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

Tagscommunity github

Works forJev for Routing

Used inJev for Software & DevTools

Languages: Python

Highlights

Ships router.py, questions.py thresholds, ask.py, smoke.py, review.py, and config.json.
Bundled examples cover pricing-page browse, refund DM, deploy retry loop, and done search.
Shadow mode with record and review.py compares Jev advice against actual bot behavior.
Truncates long fields to a 32k-token budget and recommends capped keys.

Watch out

Grok Bot obeying the route is instruction-following, not enforcement; the shared-computer key is visible to every bot on the account. 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

ai-chess-lab

Browser chess lab where LLMs, System One models, and humans play single games, Swiss tournaments, and leagues with Stockfish analysis and cost tracking.

Jev Integrations & Adapters#community#github

Community

dperezcabrera/ai-chess-labPython

systemone

Compatibility server exposing the /v1/systemone typed-decision API on top of vLLM, SGLang, or OpenRouter models, with vision models enabling image inputs that Jev and Laya lack.

Jev Integrations & Adapters#community#github

Community

infinitylogesh/systemonePython

jev-site-auditor

Site-wide content auditor combining free deterministic regex and structural checks with one batched Jev judgment call per page.

Jev Integrations & Adapters#community#github

Community

tjkimcloud/jev-site-auditorPython

AutoGPT

187.5k

Autonomous agent platform with 187,504 stars for building and running agents from plain English or a visual builder. Its platform features first-class TypeSafe Jev decision blocks for typed routing, filtering, scoring and confidence-gated next-action dispatching.

Jev Integrations & Adapters#community#autonomous-agents#platform

Community

Significant-Gravitas/AutoGPTPython

LiteLLM

60.9k

LiteLLM's proxy adds TypeSafe guardrail hooks and a Jev complexity classifier for policy judgments and request routing across 100+ model providers.

Jev Integrations & Adapters#community#python#gateway

Community

BerriAI/litellmPython

AgentScope

33k

Alibaba's multi-agent platform, which implements native TypeSafe Jev classification through JevClassifierModel for binary, choice and score routing across agent pipelines. Python 3.11+, Apache-2.0, with a FastAPI agent service and web UI.

Jev Integrations & Adapters#community#multi-agent#framework

Community

agentscope-ai/agentscopePython

Back to all resources

From the Jev community

Posts from builders shipping with Jev right now.

Follow @typesafeai