aionedge.jev-auto-router
An experimental prototype (documented in Chinese) that has Jev pick which GPT model and reasoning tier should handle each individual call within a Codex session, routed through a local proxy that keeps the session and tool loop continuous, with task correctness checked independently afterward to see whether using fewer flagship-model calls still gets the job done at lower cost.
- GitHub stars
- 5 as of 2026-09-22
- Forks
- 0
- Language
- TypeScript
- License
- Apache-2.0
- Created
- 2026-08-01
- Last updated on GitHub
- 2026-09-23
- Added to AionEdge
- 2026-09-22
What Jev does in jev-auto-router
Listed on awesomejev.com; not yet independently verified against the repository.
Verification
Integration not confirmed yet. This repository is public and listed by the community source below, but we have not yet located the code that calls Jev. Check the repository before relying on the integration, and tell us through the submission form if you find it.
Topics
- ai-agents
- ai-coding
- codex
- developer-tools
- gpt
- jev
- jev-ai
- jev-auto-router
- jev-router
- llm-router
- llm-routing
- model-routing
- openai-codex
- per-call-routing
How Community (New) projects use Jev
Recently discovered community projects that mention Jev. They were found through public lists and submissions and are waiting for a closer look, so treat each one as a lead worth exploring rather than a confirmed integration.
Related Jev projects in Community (New)
jev-chat-windows-deepseek-jevaimark-dai/jev-chat-windows-deepseek-jevA WeChat reply-assistant that reads the current chat screen via local OCR, drafts three candidate replies with DeepSeek, and can optionally route them through TypeSafe's Jev for structured review and ranking before a human sends one. An unofficial fork building on an existing Windows-only tool, adding an early macOS preview.★ 5
Python- llm-to-jevalexwestco/llm-to-jevA conversion tool that scans prompts originally written for GPT, Claude, or Gemini, finds the bounded decisions hiding inside them, and proposes equivalent Jev Choice/Score/Noul question definitions — aimed at classification-style prompts.★ 5
JavaScript - claude-jevbuchmark/claude-jevA Claude Code plugin that turns questions about your code — is this review finding real, is this debugging hypothesis worth pursuing, which of these design options is better — into typed Jev queries, giving Claude a second, independently-calibrated opinion instead of just another guess from the same model.★ 5
TypeScript - jevonlybuluoray/jevonlyA browser-automation demo that uses no language model at all: the code builds every clickable option from the page itself, and Jev's job is only to pick among them, which turns out to be enough to complete tasks like filling out a flight search or editing a Wikipedia page.★ 5
Python 
jevmojicheeaun/jevmojiA tiny web tool that takes up to 40 characters of text and returns emoji matched to it, each scored 0 to 3 by Jev rather than picking just one best guess.★ 5
JavaScript- jevchesschoxos/jevchessA no-signup web app where Jev, TypeSafe's System One model, plays chess against Stockfish, any OpenRouter-hosted LLM, or a human — showing each move's probability live and keeping a saved history of past games and win rates.★ 5
JavaScript