madewithlaya

Catalog / LLM routing & guardrails

0070Skills

Laya Router Skill: local decision routing for Codex, Claude Code, OpenCode and Pi

Cross-platform Agent Skill plus MCP tool that gives coding agents a local Laya decision capability. Verified on Apple Silicon with MLX and PyTorch backends.

Open source ↗ github.comcostfree, localtime-
wangmiaozero/laya-router-skillREADME ↗
# Laya Router Skill

[简体中文](README.zh-CN.md) · [繁體中文](README.zh-TW.md)

Universal, cross-platform local decision routing for AI coding agents, powered by Laya and Laya-MLX.

For ChatGPT Desktop (Codex), Codex CLI, Claude Code, OpenCode, Pi, and other Agent Skills compatible agents. Laya Router becomes available as a local decision capability. An agent invokes it only when its own skill or tool selection policy chooses to; use the CLI or configured MCP tool for an explicit call.

**Version: v0.2.0.** Real MLX, PyTorch and stdio MCP smoke tests passed on an Apple Silicon Mac. Explicit ChatGPT Desktop (Codex), Codex CLI, Claude Code and OpenCode calls were verified. Native Windows/Linux and Intel Mac inference remain unverified.

Laya Router is advisory only. Supported agents can invoke it when their skill or tool selection policy decides it is useful. Explicit invocation results are shown below; implicit invocation is agent-dependent and not guaranteed.

## Compatibility

| Platform | Backend | CI | Real inference |
| --- | --- | --- | --- |
| macOS Apple Silicon | laya-mlx / MLX GPU | Python 3.11/3.12 verified | Verified |
| macOS Apple Silicon (Torch) | upstream Laya / PyTorch MPS + CPU | Python 3.11/3.12 verified | Verified |
| Windows | upstream Laya / PyTorch CPU/CUDA | Python 3.11/3.12 verified | Not yet verified on a native Windows device |
| Linux | upstream Laya / PyTorch CPU/CUDA | Python 3.11/3.12 verified | Not yet verified on a native Linux device |
| Intel macOS | upstream Laya / PyTorch | Architecture covered; no native-device CI | Not verified |

| Client | Skill discovery | Explicit invocation | Implicit invocation | MCP / CLI | Status |
| --- | --- | --- | --- | --- | --- |
| ChatGPT Desktop (Codex) | Verified | Verified: three `laya_decide` calls | Agen

Also filed under LLM routing & guardrails

  1. 0100

    laya-jev-lab: a local-first cascade that matches Jev's accuracy at 1.8x the speed

    Independent Jev vs Laya measurements plus a cascade: Laya decides first, high confidence executes, low confidence escalates to Jev. At a 0.60 threshold it matched Jev's 78%.

    @yibie · LLM routing & guardrails · lower than pure Jev

  2. 0096

    Laya Ultrafast: browser-use's jev-ultrafast agent ported to local Laya on MLX

    A port of the jev-ultrafast browser agent that decides with laya-mlx instead of the hosted Jev API. DOM snapshot, executor, safety checks and inspector are the originals.

    @ipenywis · LLM routing & guardrails · free, local

  3. 0088

    laya_router: a model router with Laya matches GPT-5 nano's accuracy, 35x faster, free

    Two routers behind one endpoint, measured on 180 labelled requests. Laya answers small/medium/powerful routing in one forward pass, against a GPT-5 nano router.

    @glukicov · LLM routing & guardrails · free, local

  4. 0076

    laya-browser-agent: browser decisions on Laya, drop-in for jev-ultrafast tooling

    Local decider for the browser-agent loop: hand it a numbered table of page controls and it picks the operation and element. Playwright/CDP driver, speaks /v1/systemone.

    @ChenneyZhuang · LLM routing & guardrails · free, local, no screenshots