pi-alysis

Alysis Code Pro LLM provider and integration package for Pi. Seamlessly connects Pi to the Alysis Code gateway with native RFC 8628 OAuth device login, dynamic model discovery, and live usage tracking.

About Alysis Code Pro

Alysis Code is an AI inference gateway optimized for developers, autonomous coding agents, and high-throughput workflows.

Through its Alysis Code Pro service, it provides cost-effective, high-speed endpoints for state-of-the-art DeepSeek reasoning models. Key capabilities include:

• Massive Context Windows: 1M tokens across models for comprehensive codebase indexing, large refactors, and long-session persistence.
• Hardware-Accelerated Reasoning: Native support for DeepSeek chain-of-thought thinking streams and token-efficient tool calling.
• Seamless RFC 8628 Device Flow: One-step browser authorization without copying or managing long-lived secret keys manually.
• Zero Client-Side Metering Overhead: Server-side accounting with proactive balance alerts and real-time TUI status bar reporting.

To use this extension, sign in with your Alysis Code account via /login alysis or visit alysiscode.com.

Features

• Device Flow Authentication (RFC 8628): Run /login alysis directly in Pi to open your browser at https://alysiscode.com/activate, approve your single-use code, and persist your gateway key securely in Pi's credential store (~/.pi/agent/auth.json).
• Dynamic Model Discovery: Automatically queries the Alysis gateway API (/models) on startup to register available models without requiring package updates.
• Live Session Usage Tracking: Real-time monitoring of prompt, completion, and cache tokens with dynamic TUI status bar updates (e.g. Alysis: 14.2k tok (1.8k out)).
• Credit & Quota Alerts: Intercepts HTTP 402 (Payment Required) and 429 status codes with immediate, actionable links to recharge credits.
• Dedicated Slash Commands:
  • /alysis: Status dashboard showing gateway connectivity, active model, token statistics, and account links.
  • /alysis-usage: Granular breakdown of token consumption and turn counts.
  • /alysis-models: Runtime refresh and catalog synchronization from the gateway.
  • /logout alysis: Instant session logout and credential revocation.
• Resilient Fallback: Instant startup with built-in model definitions if offline or before authentication is configured.

Supported Models

Models are dynamically retrieved from the gateway and configured with a 1M token context window, 16,384 max output tokens, DeepSeek reasoning format, and zero client-side cost deduction:

Installation

Install pi-alysis globally in Pi using either npm or GitHub:

From npm

pi install npm:pi-alysis

From GitHub

pi install git:github.com/jh0nz/pi-alysis

Temporary Run (without installation)

pi -e npm:pi-alysis

Local Development / Project Install

To install for the current project only:

pi install -l /path/to/pi-alysis

Authentication

You can authenticate using Pi's interactive /login command or by setting an environment variable.

Option 1: Native Pi Login (Recommended)

Start Pi and run the login command:

/login alysis

Pi will display your one-time code (e.g. ABCD-EFGH) and open your browser at https://alysiscode.com/activate?code=ABCD-EFGH. Approve the code on the web page, and Pi will securely store your credentials in ~/.pi/agent/auth.json.

Option 2: Environment Variable

Export your gateway key in your shell configuration (~/.bashrc, ~/.zshrc, or ~/.config/fish/config.fish):

export ALYSIS_API_KEY="slk_your-gateway-key-here"

Optional Custom Endpoints

Usage

Interactive Selection

Start Pi and open the model selector:

/model

Select any alysis/... model from the list.

Direct Launch

# DeepSeek V4 Flash (1M context window, fast reasoning)
pi --model alysis/deepseek-v4-flash

# DeepSeek V4 Pro (1M context window, deep reasoning)
pi --model alysis/deepseek-v4-pro

# DeepSeek V4 Flash Vision
pi --model alysis/deepseek-v4-flash-vision-exp

List Available Models

pi --list-models | grep alysis

Session Commands

Inside Pi, manage your session with dedicated slash commands:

• /alysis — View connection state, active model, and token metrics.
• /alysis-usage — View detailed token metrics and session summary.
• /alysis-models — Refresh available models dynamically from the gateway.
• /logout alysis — Disconnect and clear credentials.

Technical Details

• Base URL: https://vzigujbcjjmpntxhmyvr.supabase.co/functions/v1/llm/v1
• Protocol: High-performance openai-completions engine with Server-Sent Events (SSE) streaming.
• Authentication: OAuth 2.0 Device Authorization Grant (RFC 8628) with polling and verification URI.
• Dynamic Endpoint: Queries GET /models with a 3.5s timeout safeguard on startup.
• Status Bar Integration: Custom footer listener hooked into agent events for live token tracking.

Project Structure

pi-alysis/
├── src/
│   ├── constants.ts   # Endpoint definitions, public anon key, and env helpers
│   ├── auth.ts        # RFC 8628 device flow implementation
│   ├── models.ts      # Dynamic /models discovery and model mapping
│   ├── usage.ts       # Session token tracking and footer formatting
│   └── index.ts       # Extension entry point, event listeners, and slash commands
├── dist/              # Compiled ESM bundle and TypeScript declarations
├── package.json       # Pi package manifest
├── tsconfig.json      # TypeScript compiler configuration
└── README.md          # Documentation and setup guide

License

MIT (c) jh0nz