@wishx127/pi-fast

Fast and Ultrafast toggles for supported models in Pi

Packages

Package details

extension

Install @wishx127/pi-fast from npm and Pi will load the resources declared by the package manifest.

npm repo home report

$ pi install npm:@wishx127/pi-fast

Package: @wishx127/pi-fast
Version: 0.1.3
Published: Oct 9, 2026
Downloads: 684/mo · 208/wk
Author: ny9u
License: MIT
Types: extension
Size: 11.5 KB
Dependencies: 0 dependencies · 1 peer

Pi manifest JSON

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

Security note

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

README

pi-fast

Use /fast or /ultrafast to request faster processing for supported models in Pi. The commands change the service tier while preserving your model and reasoning level.

简体中文

Quick start

Try it for one Pi run without installing it permanently:

pi --no-extensions -e npm:@wishx127/pi-fast

To install it permanently:

pi install npm:@wishx127/pi-fast

In Pi, type /fast or /ultrafast. Both commands take no arguments. The modes are mutually exclusive and start off by default:

Enabling a mode requires the current model and provider to support it. If unsupported, the command warns you and leaves both the active and saved modes unchanged. For example, requesting Ultrafast on a model that only supports Fast keeps Fast active. Turning off the currently selected mode always works.

What it does

When enabled, the extension checks the Pi provider, active model, and serialized request model. It adds service_tier: "priority" for Fast or service_tier: "ultrafast" for Ultrafast before Pi sends the request, unless a tier is already set. OpenAI documents priority as the backward-compatible alias for Fast mode. Other requests are unchanged. The extension does not make independent network requests or alter your prompts, tools, model, or reasoning level.

Provider	Fast model IDs	Ultrafast model IDs
`openai` (API key)	`gpt-5.4`, `gpt-5.4-mini`, `gpt-5.5`, `gpt-5.6-sol`, `gpt-5.6-terra`, `gpt-5.6-luna`, `gpt-6-astra`, `gpt-6.1-sol`	`gpt-6-astra`, `gpt-6.1-sol`
`openai-codex` (ChatGPT sign-in)	`gpt-5.4`, `gpt-5.5`, `gpt-5.6-sol`, `gpt-5.6-terra`, `gpt-5.6-luna`, `gpt-6-astra`, `gpt-6-sol`, `gpt-6-luna`, `gpt-6.1-sol`	`gpt-6-astra`, `gpt-6.1-sol`

Important: requested ≠ delivered

Command notifications report the locally requested mode, not the tier actually used. OpenAI may reject or downgrade a request; check the response or provider usage dashboard. Existing tiers set by other configuration or extensions take precedence.

Contributing

Contributions are welcome! Open an issue to report a bug or discuss an idea, or submit a pull request with a fix. When adding model support, link the provider's documentation for that mode and include a test.

MIT licensed.