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

Session-scoped Codex subscription Fast and Ultrafast modes for Pi requests.

Packages

Package details

extension

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

npm repohome report
$ pi install npm:@firstpick/pi-extension-codex-fast-mode
Package
@firstpick/pi-extension-codex-fast-mode
Version
0.1.2
Published
Oct 1, 2026
Downloads
271/mo · 220/wk
Author
firstpick
License
MIT
Types
extension
Size
21 KB
Dependencies
0 dependencies · 1 peer
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

Codex Fast Mode for Pi

Choose Normal, Fast, or Ultrafast processing for subscription-backed Codex sessions.

What you can do

  • Select Fast mode or GPT-6 Astra Ultrafast for the current session branch.
  • Check your selected mode with /fast-mode status.
  • Keep your preference when resuming or navigating that branch.
  • Leave other providers and API-key billing unchanged.

Install

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

Restart Pi or run /reload after installing or updating the extension.

How to use it

Select a subscription-backed Codex model, then run /fast-mode on to request Fast processing. Run /fast-mode status to inspect the preference, or /fast-mode off to stop overriding the request tier.

For Ultrafast, select GPT-6 Astra through the openai-codex provider, check the requirements below, then run:

/fast-mode ultrafast
/fast-mode status
  • /fast-mode switches Normal to Fast, or turns either enabled mode off.
  • /fast-mode on or /fast-mode fast selects Fast.
  • /fast-mode ultrafast selects Ultrafast for GPT-6 Astra.
  • /fast-mode off or /fast-mode normal removes the override.
  • /fast-mode status reports the current session-branch preference.

The updated Pi Web UI companion also offers Normal, Fast, and Ultrafast under Codex Usage. It asks you to confirm the higher cost before selecting Ultrafast.

Before you start

Ultrafast uses substantially more allowance or credits. It currently requires Pro $500 or eligible Enterprise/Edu access and uses 8× Standard included usage or 6× purchased-credit/pay-as-you-go usage. Workspace terms and permissions may differ. Other self-serve plans do not have subscription Ultrafast access at launch, even with purchased credits.

Fast also increases usage. Selecting a mode requests that tier; it does not prove OpenAI accepted it. The extension does not inspect your plan or credentials, choose a model, or add missing models. Changes are blocked while Pi is busy.

Technical details

See TECHNICAL.md for all commands, eligibility, credit use, migration, and troubleshooting.