pi-gpt-service-tier

Toggle OpenAI Responses priority service tier per pi session

Package details

extension

Install pi-gpt-service-tier from npm and Pi will load the resources declared by the package manifest.

$ pi install npm:pi-gpt-service-tier

Pi manifest JSON

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

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

Provides the /gpt-fast command for pi, allowing OpenAI Responses API Priority processing to be toggled per session.

Disabled by default for new sessions; service_tier is omitted from request bodies.
After entering /gpt-fast for the first time, requests whose api is openai-responses include service_tier: "priority".
Entering /gpt-fast again removes service_tier from those requests.
Other API types, such as openai-completions, are not modified.
The setting is stored in the current session and restored when that session is resumed.
When enabled, • fast appears after the current model in the pi TUI footer; it is removed when disabled.

Install dependencies and run the full check suite:

npm install
npm run check

Load the extension directly during development:

pi -e ./extensions/index.ts

Run this from the repository directory:

pi install .

After publishing to npm, install it by package name:

pi install npm:pi-gpt-service-tier

Start pi after installation, then enter /gpt-fast to toggle the setting. The command reports whether it is currently enabled or disabled.

Inspect the package contents before publishing:

npm pack --dry-run
npm publish

MIT