pi-commandcode-pricing

Command Code plan pricing in pi — /go-pricing, /goat-pricing, /pro-pricing, /max-pricing popups scraped live from commandcode.ai

Packages

Package details

extension

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

npm repohome report
$ pi install npm:pi-commandcode-pricing
Package
pi-commandcode-pricing
Version
0.1.0
Published
Sep 5, 2026
Downloads
192/mo · 9/wk
Author
lemonphase
License
MIT
Types
extension
Size
29.1 KB
Dependencies
0 dependencies · 2 peers
Pi manifest JSON
{
  "extensions": [
    "./extensions"
  ]
}

Security note

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

README

pi-commandcode-pricing

Command Code plan pricing inside pi — scrollable popup with sort, search, and request limits, scraped live from the official docs pages (no public pricing API exists).

Install

pi install npm:pi-commandcode-pricing
# or from GitHub:
pi install git:github.com/LemonPhase/pi-commandcode-pricing

Or try without installing:

pi -e git:github.com/LemonPhase/pi-commandcode-pricing

Commands

Command Shows
/go-pricing Go plan — 40 Go-eligible models with per-token prices (windows aren't published per-model on the Go page)
/goat-pricing GOAT plan — per-model pricing, intel, request limits (free models included)
/pro-pricing Pro plan — 46 models across three credit tables
/max-pricing Max 10× / Max 20× — both credit tiers side by side (they differ only in credit amounts)

Keys inside a popup: Tab cycles sort (Credits → Intelligence → Plan intel/mo), / filters by model name (Backspace on empty cancels), ↑↓/jk scroll, ←→/hl pan columns when overflowing, PgUp/PgDn/u d page, g/G top/end, Esc clears search then closes.

Where the data comes from

  • Each command reads its own plan page (commandcode.ai/docs/plans/<plan>) for prices, credits, and request windows.
  • Intelligence scores always come from the GOAT page's embedded flight JSON, which covers the whole model catalogue — the same scores on every plan.
  • Value sort ranks by intel per blended $/MTok (0.75·input + 0.25·output, matching the site's own calculator); free models pin to the top. That's the plan-blind API-price lens.
  • Intel/mo is the plan-holder's lens: intel × the plan's monthly credit allowance for that model ÷ blended cost — the scored intelligence×MTok your plan buys per month. Credits differ per model, so this sees what Intel/$ can't. Max's two tiers show joined (448/897); Go shows — (no per-model allowances published); free models show ∞. Plan sort ranks by it.
  • No caching — every command fetches fresh. Pages are ~300–650KB; expect a short delay.

Development

npm install
./test/fetch-fixtures.sh   # refresh saved plan-page fixtures
node test/run.mjs          # headless checks against fixtures

To load a local checkout into pi without installing:

pi install /path/to/pi-commandcode-pricing   # local-path package, loads live from disk

License

MIT