pi-live-codex

Realtime gpt-live-1-codex voice frontend for the Pi coding agent

Packages

Package details

extension

Install pi-live-codex from npm and Pi will load the resources declared by the package manifest.

npm repo home report

$ pi install npm:pi-live-codex

Package: pi-live-codex
Version: 0.1.10
Published: Oct 27, 2026
Downloads: 494/mo · 47/wk
Author: brettinternet
License: MIT
Types: extension
Size: 144.1 KB
Dependencies: 2 dependencies · 4 peers

Pi manifest JSON

{
  "extensions": [
    "./index.ts"
  ]
}

Security note

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

README

pi-live-codex

Talk to OpenAI Codex from Pi.

pi install npm:pi-live-codex

Requires Node 22.19+ and a microphone.

/login openai-codex
/live            start voice mode
/live <voice>    start with a voice

Key	Action
`Ctrl+L`	Toggle live mode
`Esc`	End voice mode
`Space` (empty editor)	Mute or resume
Type or paste, then `Enter`	Share text with voice without starting work

Shared text appears in the live transcript and reaches voice immediately as silent reference context. Then describe what you want aloud; the next backend request includes the text. Enter alone does not request analysis or approve an action.

Drop image files into the session to attach them to the next backend request. Voice receives the filenames, not the image contents; it can discuss the images after the backend returns its findings. Dropping images does not start work. Clipboard image paste is not supported by the live editor.

Requests made while one is running are queued, with status updates and cancellation. Voice also tracks background work such as subagents, can cancel it, and can answer confirmation requests from other extensions.

Only one Pi session owns the microphone at a time. Starting voice in another session pauses the first; it keeps its transcript and drafts, and its work keeps running.