Skip to main content

Natural language terminal agent for Windows PowerShell

Project description

Termagent: Natural Language Terminal Agent for Windows

Termagent is a powerful, Windows-native AI agent that transforms natural language into executable PowerShell commands, professional emails, research documents, web searches, and full GitHub workflows — all from a single terminal interface. Built on LangGraph with Groq inference, it combines near-instant response times with intelligent intent routing, dual-layer safety, and a stunning Textual TUI.

🚀 Key Features

  • Natural Language → PowerShell: Convert plain English into complex PowerShell scripts instantly.
  • Smart Intent Routing: Five-way LLM classifier automatically routes to the right pipeline — command, chat, email, document, or GitHub.
  • Session Memory: Retains last few messages as history within a session for contextual follow-ups.
  • Web Search (ReAct): Answers questions with real-time information via DuckDuckGo search, synthesized through a ReAct tool-use loop.
  • Document Creation: Generates well-formatted .docx reports with headings, tables, bullet points, and inline formatting — powered by web research and LLM knowledge.
  • Email Integration: Compose and send professional emails with attachments via Gmail SMTP. Structured output handles recipient, subject, body, and multi-file attachments.
  • GitHub & Git (MCP-Powered): Full git and GitHub operations via the Model Context Protocol — commit, push, create releases, list PRs, manage issues, and more. Combines local git tools with dynamically-discovered GitHub MCP tools in a ReAct loop.
  • Coding Tasks: Capable of handling coding tasks. Plans and executes various tasks, HITL for safety.
  • File Reading: Read .docx, .pdf, .txt, and .md files for use in chat or document workflows.
  • Dual-Layer Safety: Static blacklist blocks dangerous patterns instantly; LLM security review catches context-sensitive risks.
  • Human-in-the-Loop (HITL): Risky commands always require explicit user confirmation before execution.
  • Raw PowerShell Mode: Prefix with ! to bypass the agent and run PowerShell directly.
  • Textual TUI: Dark-themed terminal UI with animated spinner, rich markdown rendering, and status indicators.

🛠️ Architecture

Termagent uses a LangGraph state machine with pre-check short-circuiting, five-way intent classification, and multiple ReAct loops for tool-augmented nodes. Inference runs on Groq (llama-3.3-70b for classification, gpt-oss-120b for reasoning) for near-instant responses.

🚀 Quick Start

pip install termagent-cli
termagent

⚠️ Safety First

Termagent executes real commands on your system. Always review flagged commands before confirming.

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

termagent_cli-1.1.0.tar.gz (32.1 kB view details)

Uploaded Source

Built Distribution

If you're not sure about the file name format, learn more about wheel file names.

termagent_cli-1.1.0-py3-none-any.whl (30.5 kB view details)

Uploaded Python 3

File details

Details for the file termagent_cli-1.1.0.tar.gz.

File metadata

  • Download URL: termagent_cli-1.1.0.tar.gz
  • Upload date:
  • Size: 32.1 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/6.2.0 CPython/3.13.11

File hashes

Hashes for termagent_cli-1.1.0.tar.gz
Algorithm Hash digest
SHA256 bcb3bd8c9a3879e6320d26a8016109d9777de8c621a62798570849e8df868f98
MD5 9cecf505fceddeeb3b122891ebc5fa73
BLAKE2b-256 0cdf6551211468a0494c3ac0db86253a00e73534e0abdf2964351052bfd6b6e0

See more details on using hashes here.

File details

Details for the file termagent_cli-1.1.0-py3-none-any.whl.

File metadata

  • Download URL: termagent_cli-1.1.0-py3-none-any.whl
  • Upload date:
  • Size: 30.5 kB
  • Tags: Python 3
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/6.2.0 CPython/3.13.11

File hashes

Hashes for termagent_cli-1.1.0-py3-none-any.whl
Algorithm Hash digest
SHA256 f698fd198402dd106ce3a108e7762d63e459f00f4c86884cc6bf1ab94ee4a71c
MD5 c34f1d4bb3d7092d375a81fb7d695c34
BLAKE2b-256 513c2fab058f01b24830e185fc3fa25faddd9911ae5e95d8a2cd5c968cd93735

See more details on using hashes here.

Supported by

AWS Cloud computing and Security Sponsor Datadog Monitoring Depot Continuous Integration Fastly CDN Google Download Analytics Pingdom Monitoring Sentry Error logging StatusPage Status page