@ryan_nookpi/pi-extension-codex-fast-mode

Codex fast mode extension for pi.

Packages

Package details

extension

Install @ryan_nookpi/pi-extension-codex-fast-mode from npm and Pi will load the resources declared by the package manifest.

npm repohome report
$ pi install npm:@ryan_nookpi/pi-extension-codex-fast-mode
Package
@ryan_nookpi/pi-extension-codex-fast-mode
Version
0.2.7
Published
Sep 5, 2026
Downloads
891/mo · 151/wk
Author
ryan_nookpi
License
MIT
Types
extension
Size
8.6 KB
Dependencies
0 dependencies · 2 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

@ryan_nookpi/pi-extension-codex-fast-mode

This extension helps pi use OpenAI Codex in a faster, lower-verbosity mode.

It is intended for openai-codex with gpt-5.4 through the supported gpt-5.6 variants, plus gpt-6-astra, when you want quick execution and shorter responses.

Install

pi install npm:@ryan_nookpi/pi-extension-codex-fast-mode

Great for

  • prioritizing speed over long explanations
  • keeping Codex responses concise
  • toggling a faster Codex setup per session

Usage

/codex-fast on
/codex-fast off
/codex-fast status

Notes

  • Target models: openai-codex / gpt-5.4, gpt-5.5, gpt-5.6-luna, gpt-5.6-sol, gpt-5.6-terra, and gpt-6-astra.
  • Command descriptions and status messages abbreviate the list as gpt-5.4 ~ gpt-5.6, gpt-6-astra; matching still uses the exact model IDs above.
  • This extension does not register models. Pi must already offer openai-codex/gpt-6-astra through its model catalog or your models.json. The bundled catalog in Pi SDK 0.85.0 does not include Astra; upstream support was added after that release.
  • Astra availability depends on your account. Where available, Codex Astra Fast mode consumes credits at 2.5x the Standard rate.
  • It always applies text.verbosity=low.
  • When fast mode is enabled, it also injects service_tier=priority.
  • The setting is stored locally and persists across sessions.

References