pi-jetbrains-junie-bridge

Pi extension that connects Pi coding agent with JetBrains Junie as the AI backend

Packages

Package details

extension

Install pi-jetbrains-junie-bridge from npm and Pi will load the resources declared by the package manifest.

$ pi install npm:pi-jetbrains-junie-bridge
Package
pi-jetbrains-junie-bridge
Version
0.2.4
Published
Jul 4, 2026
Downloads
not available
Author
sschatter
License
MIT
Types
extension
Size
317 KB
Dependencies
0 dependencies · 0 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

pi-jetbrains-junie-bridge

A Pi extension that lets you use JetBrains Junie as the AI backend for the Pi coding agent — using your existing Junie subscription.

Install

pi install npm:pi-jetbrains-junie-bridge

Then inside pi run the /login command. Select Use a subscription and then JetBrains Junie to authenticate.

Once authenticated, run /model to select a model provided by the junie bridge.

Features

  • OAuth login — browser-based JetBrains authentication with automatic token refresh
  • Balance tracking — session cost and remaining balance shown in Pi's status line (see footer) Screnshot showcasing balance tracking
  • /junie command — check proxy status, balance, and available models from within Pi Screnshot showcasing the `junie` command

Available Models

Only OpenAI and Anthropic (Claude) models are supported. Google/Gemini models are not supported — the Grazie backend requires a native protocol for these and rejects the OAuth tokens used by this proxy.

Recommendation: Use the Claude models — only Anthropic models send reasoning blocks (extended thinking) to Pi, which significantly improves coding output quality.

Anthropic:

  • claude-sonnet-4-6
  • claude-sonnet-5
  • claude-opus-4-6
  • claude-opus-4-7
  • claude-opus-4-8
  • claude-fable-5

OpenAI:

  • openai-gpt-5-2
  • openai-gpt-5-4 — reasoning effort not adjustable (OpenAI limitation with tool use)
  • openai-gpt-5-5 — reasoning effort not adjustable (OpenAI limitation with tool use)

Requirements

How It Works

The extension starts a local proxy server that translates between Pi and JetBrains' Grazie backend. On first use, you'll authenticate via JetBrains OAuth (browser-based PKCE flow).

┌─────────┐     ┌──────────────┐     ┌──────────────┐
│   Pi    │────>│  pi-junie    │────>│ Junie/Grazie │
│ agent   │     │  local proxy │     │   backend    │
└─────────┘     └──────────────┘     └──────────────┘
  • OpenAI models (openai-gpt-*) are forwarded via /v1/chat/completions
  • Anthropic models (claude-*) are forwarded via /v1/messages
  • The proxy runs on an ephemeral port and shuts down when Pi exits

Disclaimer

This extension includes a proxy server reverse-engineered from the official JetBrains Junie CLI. It is not officially supported by JetBrains. Use it in accordance with the JetBrains AI Service Terms of Service.

License

MIT