microsoft/intelligent-terminal

What it solves

Intelligent Terminal addresses the friction of switching between a command-line interface and an AI assistant. It integrates AI agents directly into the terminal environment, eliminating the need to copy-paste shell output or manually provide context to an LLM for troubleshooting and command generation.

How it works

It is an experimental fork of Windows Terminal that acts as a local transport layer. It uses the Agent Client Protocol (ACP) to communicate with agent CLIs installed on the user's machine (such as GitHub Copilot CLI, Claude, or Gemini). The terminal captures shell output and provides it as context to the selected agent, which can then suggest commands that the user must explicitly approve before execution.

Who it’s for

Developers and system administrators using Windows who want AI-assisted command-line workflows, automated error diagnosis, and the ability to run local or cloud-based AI models within their terminal.

Highlights

• Agent Pane: A docked, context-aware chat interface that can spin up background tasks in new tabs.
• Error Detection: Automatically detects command failures and provides a one-click option to get AI-generated fix suggestions.
• BYOK (Bring Your Own Key): Supports custom providers and local models (e.g., via Ollama) using OpenAI-compatible endpoints.
• Command Palette Integration: Allows users to trigger agent tasks directly from the command palette using a ? prefix.
• Multi-Agent Support: Ability to pin specific agents to different shell profiles (e.g., a specific agent for Ubuntu/WSL).