@neuron-mr-white/pi-zenmux

ZenMux provider extension for Pi — dynamic model discovery with caching

Package details

extension

Install @neuron-mr-white/pi-zenmux from npm and Pi will load the resources declared by the package manifest.

$ pi install npm:@neuron-mr-white/pi-zenmux
Package
@neuron-mr-white/pi-zenmux
Version
1.0.1
Published
Apr 25, 2026
Downloads
169/mo · 169/wk
Author
neuron-mr-white
License
MIT
Types
extension
Size
6.8 KB
Dependencies
0 dependencies · 1 peer
Pi manifest JSON
{
  "extensions": [
    "./extensions/index.ts"
  ]
}

Security note

Pi packages can execute code and influence agent behavior. Review the source before installing third-party packages.

README

pi-zenmux

ZenMux provider extension for Pi. Auto-discovers and registers all available models from the ZenMux API gateway.

Install

pi install npm:@neuron-mr-white/pi-zenmux

Setup

Set your ZenMux API key:

export ZENMUX_API_KEY="your-api-key"

Usage

Start pi normally — ZenMux models appear under the zenmux provider:

pi

Select a model with /model or use zenmux/<model-id> directly.

Commands

  • /zenmux-refresh — Force refresh model list from API

How it works

On startup, the extension:

  1. Reads cached model list from ~/.cache/pi-zenmux/models.json (24h TTL)
  2. If cache hit: registers models synchronously (zero latency)
  3. If cache miss: registers empty provider, fetches models in background
  4. Writes cache for next startup

Development

Test locally without publishing:

ZENMUX_API_KEY="your-key" pi -e ./extensions/index.ts