aiterm

AiTerm

Say hello to AiTerm, your new best friend for chatting with AI right from the terminal! This nifty command-line tool lets you toss natural language commands like "list files" into a slick text-based interface, powered by tview. Watch as the AI whips up the perfect terminal command (like ls) and runs it in a tmux split pane, all while keeping your conversation history scrollable and snappy. Whether you're a terminal ninja or just love a good chat, AiTerm makes AI-powered workflows a breeze!

Features

• Natural Language Interaction: Enter commands in natural language (e.g., "list files"), and the AI generates corresponding terminal commands (e.g., ls) and explanations.
• tmux Integration: Commands are executed in a dynamically created tmux pane, displayed alongside the chat interface.
• Scrollable Chat History: View past conversations in a scrollable TextView, with support for manual scrolling to review history.
• Dynamic "Generating" Indicator: A visual animation in the input field ("generating.", "generating..", "generating...") shows when the AI is processing.
• Cancel Generation: Press Ctrl+C during AI processing to cancel and resume input.
• Configuration Management: Stores AI service details (URL, Token, Model) in ~/.aitermrc, with interactive terminal-based setup on first run.
• Version Information: Use aiterm -v to display the tool's version.
• Error Handling: Validates AI configuration and handles tmux session requirements gracefully.

Installation

Install AiTerm using Go:

go install github.com/aki-colt/aiterm@latest

Ensure $GOPATH/bin is in your PATH:

export PATH=$PATH:$(go env GOPATH)/bin

Requirements

• Go: 1.22 or later
• tmux: Required for pane splitting and command execution
• Terminal: A terminal emulator supporting ANSI colors and keyboard input
• Platform: AiTerm only supports for unix-like system such as linux or macos for now.

Quick Start

1. Start a tmux session:

   tmux
   

2. Run the tool:

   aiterm
   

3. Configure AI (first run only): If ~/.aitermrc is missing or incomplete, you'll be prompted to enter:

   Enter AI URL: https://api.example.com/v1
   Enter API Token: your-token
   Enter Model: gpt-4
   

   This creates ~/.aitermrc:

   [ai]
   url=https://api.example.com/v1
   token=your-token
   model=gpt-4
   

4. Interact:

   • Type natural language commands (e.g., "list files") in the input field.
   • Press Enter to send the command to the AI.
   • View AI responses and command outputs in the split tmux pane.
   • Use Ctrl+C to cancel AI processing.
   • Press Ctrl+Q to exit.

5. Check version:

   aiterm -v
   

   Output:

   aiterm version v0.1.0
   

Configuration

The tool stores AI configuration in ~/.aitermrc locolly. Edit this file directly to update settings, or delete it to trigger the configuration prompt on next run.

Example ~/.aitermrc:

[ai]
url=https://api.example.com/v1
token=your-token
model=gpt-4

For Contributors

Project Setup

1. Clone the repository:

   git clone https://github.com/yourusername/aiterm.git
   cd aiterm
   

2. Install dependencies:

   go mod tidy
   

3. Build and test:

   go build -o aiterm
   ./aiterm
   

Project Structure

aiterm/
├── README.md
├── ai
│   ├── ai.go # request to llm
│   ├── prompt.go # prompt
│   └── tools.go # tools to check and execute commands
├── cmd
│   └── main.go # entry point, handles flags, config, and UI setup
└── terminal
│   ├── command.go # cotroller of tmux
│   ├── dialog.go # dialog ui
├── go.mod
├── go.sum

Implementation Overview

AiTerm is built with the following components:

• Command-Line Interface:

  • Uses flag package to handle -v for version display.
  • Checks ~/.aitermrc for AI configuration (URL, Token, Model).
  • Prompts for configuration via terminal input if missing, validated using openai-go SDK.

• Configuration:

  • Stores settings in ~/.aitermrc using ini format.
  • Validates AI config by attempting to list models with openai-go.

• UI:

  • Built with tview, featuring a TextView for chat history and an InputField for user input.
  • Supports scrolling (SetScrollable(true)), dynamic "generating" animation, and Ctrl+C cancellation.
  • Updates are queued via app.QueueUpdateDraw for thread safety.

• tmux Integration:

  • Managed by tmux.go, which splits the terminal using tmux split-window -P.
  • Sends commands to the correct pane using send-keys, tracked via pane ID.

• AI Processing:

  • Uses openai-go SDK to process input.
  • Runs in a goroutine with context for cancellation support.
  • Displays results in TextView and executes commands in tmux.

Contribution Ideas

• Add support for multiple AI providers (e.g., Anthropic, Grok).
• Implement command-line flags for configuration overrides.
• Enhance UI with color-coded messages or a status bar.
• Add unit tests for configuration and tmux logic.
• Support configuration via environment variables.
• Support windows system.

License

MIT License