High-agency PowerShell AI framework for multi-agent orchestration and autonomous systems engineering. PSAI bridges systems engineering and LLMs, transforming AI into a functional component within PowerShell. It empowers developers and architects to build, test, and deploy autonomous agents that interact with consoles, file systems, and enterprise APIs, moving AI beyond chatbots to robust engineering tools.

How It Works

PSAI facilitates autonomous multi-step task execution by orchestrating PowerShell functions as LLM-callable "Tools," adhering to Model Context Protocol (MCP) patterns and established systems engineering disciplines. It provides state and memory management for complex, long-running conversations, including rewind and message listing. The framework's provider-agnostic design supports OpenAI, Azure OpenAI, and Anthropic Claude, enabling seamless integration across various LLM services.

Quick Start & Requirements

Installation is performed via PowerShell: Install-Module -Name PSAI. Users must configure API keys for their chosen LLM provider (OpenAI or Azure OpenAI), typically by setting environment variables like $env:OpenAIKey or configuring Azure secrets using Set-AzOAISecrets.

Highlighted Details

• Autonomous Agents: Create interactive tools powered by LLMs for complex problem-solving, integrating AI directly into terminal workflows.
• Extensible Tooling: Register any PowerShell function as a tool, enabling LLMs to securely interact with the local environment.
• ConvertTo-AIPrompt: A utility to package GitHub repositories into an AI-optimized XML format for easier LLM ingestion.
• Slash Commands: Interactive session commands like /clear, /rewind, and /listmsgs enhance user control within agent conversations.
• Invoke-QuickPrompt: Supports passing custom PowerShell functions as tools and visualizing tool calls within responses.

Maintenance & Community

Feedback and questions can be directed via GitHub issues, Twitter, or LinkedIn. Specific community links or contributor details were not explicitly provided in the README.

Licensing & Compatibility

The specific open-source license is not explicitly stated in the README. The framework is compatible with OpenAI, Azure OpenAI, and Anthropic Claude LLM providers and integrates directly into the PowerShell ecosystem.

Limitations & Caveats

The project outlines a future roadmap including additional tools (weather, document parsing) and richer CLI interactions, indicating ongoing development. Users are responsible for managing LLM API key security and associated costs. The absence of a stated license may present adoption risks for commercial or sensitive projects.