AI Commander for WezTerm

Integrates with AI providers (Anthropic Claude and OpenAI GPT) to generate and select bash commands based on natural language prompts.

Features

• Command Generation: Generate multiple bash command options from natural language descriptions
• Interactive Selection: Choose from generated commands using a selection menu
• Context-Aware Generation: Select text in the terminal to provide context for command generation
• Prompt History: Access and reuse previous prompts with Alt+Shift+H
• Quick Command Generation: Generate commands on-the-fly with Alt+Shift+X
• Word Deletion: Enhanced text editing with Alt+Backspace for word deletion

Prerequisites

1. API Key: Get your API key from:
   • Anthropic Console for Claude
   • OpenAI Platform for GPT
2. WezTerm: Recent version with Lua plugin support

Installation

1. Add the plugin to your .wezterm.lua configuration:

   local wezterm = require 'wezterm'
   local config = wezterm.config_builder()
   local act = wezterm.action
   
   -- Load AI Commander plugin
   local ai_plugin = wezterm.plugin.require("https://github.com/dimao/ai-commander.wezterm")
   
   -- Apply AI Commander configuration with your API key
   ai_plugin.apply_to_config(config, {
     provider = "anthropic",  -- "anthropic" or "openai"
     api_key = {
       anthropic = "your-anthropic-api-key-here",  -- Replace with your actual API key
       openai = "your-openai-api-key-here"        -- Replace with your actual API key
     }
   })
   
   -- Your existing keybindings
   config.keys = {
     -- Your other keybindings here...
     
     -- AI Commander keybindings (add these to your existing keys table)
     {
       key = 'Backspace',
       mods = 'ALT',
       action = act.SendKey { key = 'w', mods = 'CTRL' },
     },
     {
       key = 'H',
       mods = 'ALT|SHIFT',
       action = wezterm.action_callback(function(window, pane)
         ai_plugin.show_history(window, pane)
       end),
     },
     {
       key = 'X',
       mods = 'ALT|SHIFT',
       action = wezterm.action_callback(function(window, pane)
         ai_plugin.show_prompt(window, pane)
       end),
     },
   }
   
   return config

Usage

The plugin provides the following keybindings (which you need to add to your config):

• Alt+Shift+X: Generate bash commands from a prompt
• Alt+Shift+H: Access prompt history
• Alt+Backspace: Delete word (enhanced text editing)

Note: The keybindings are not automatically registered by the plugin. You must add them to your config.keys table as shown in the installation example above.

Workflow

Basic Command Generation

1. Press Alt+Shift+X to open the command generation prompt
2. Enter a natural language description of what you want to do (e.g., "find all PDF files in current directory")
3. The AI will generate multiple bash command options (configurable, default: 5)
4. Select the command you want to use from the interactive menu
5. The selected command will be inserted into your terminal

Context-Aware Command Generation

1. Select text in your terminal (drag to highlight text like file paths, error messages, log entries, etc.)
2. Press Alt+Shift+X to open the command generation prompt
3. The prompt description will indicate that selected text will be used as context
4. Enter your command request (e.g., "fix this error" or "process these files")
5. The AI will generate commands that take the selected text into account
6. Select and execute the appropriate command

Example Use Cases:

• Select an error message and ask "how to fix this"
• Select file paths and ask "delete these files"
• Select log entries and ask "parse this data"
• Select directory listing and ask "find largest files"

Prompt History

Access your previous prompts with Alt+Shift+H:

1. Press Alt+Shift+H to see your prompt history
2. Select a previous prompt from the list
3. The AI will regenerate commands for that prompt

Configuration

You can customize the plugin by passing configuration options to apply_to_config:

ai_plugin.apply_to_config(config, {
  provider = "anthropic",  -- "anthropic" or "openai" (default: "anthropic")
  api_key = {
    anthropic = "your-anthropic-api-key-here",  -- Your Anthropic API key
    openai = "your-openai-api-key-here"        -- Your OpenAI API key
  },
  max_tokens = 4000,  -- Maximum response length
  temperature = 0.1,  -- Response creativity (0.0-1.0)
  command_count = 5,  -- Number of commands to generate (default: 5)
  history_file = wezterm.home_dir .. '/.wezterm_ai_prompt_history.txt',  -- History file location
  max_history = 100,  -- Maximum number of history items
})

Provider-specific Options

The plugin automatically uses the correct API endpoint and model for each provider:

Anthropic (Claude):

• API URL: https://api.anthropic.com/v1/messages
• Default model: claude-3-5-sonnet-20241022

OpenAI (GPT):

• API URL: https://api.openai.com/v1/chat/completions
• Default model: gpt-4

You can override these by setting custom values:

ai_plugin.apply_to_config(config, {
  provider = "openai",
  model = {
    anthropic = "claude-3-5-sonnet-20241022",
    openai = "gpt-4o"  -- Use GPT-4o instead of GPT-4
  },
  api_url = {
    anthropic = "https://api.anthropic.com/v1/messages",
    openai = "https://api.openai.com/v1/chat/completions"
  },
  -- ... other options
})

How It Works

1. Prompt Processing: Your natural language input is enhanced with a specific instruction to generate bash commands
2. AI Generation: The selected AI provider (Claude or GPT) generates 3-5 different command options based on your description
3. Command Parsing: The response is parsed to extract individual commands
4. Selection Interface: If multiple commands are generated, you get an interactive menu to choose from
5. History Management: Prompts are automatically saved to a history file for reuse

Configuration Summary

• provider: Choose between "anthropic" (Claude) or "openai" (GPT)
• api_key: Your API keys for the respective providers
• command_count: Number of commands to generate (default: 5)
• max_tokens, temperature: Control AI response behavior
• history_file, max_history: Manage prompt history storage

Files

• ~/.wezterm_ai_prompt_history.txt: Stores your prompt history (automatically managed)

Troubleshooting

1. "[provider] API key not configured"

   • Ensure you've set the appropriate api_key.[provider] parameter in your plugin configuration
   • Check that your API key is valid and properly formatted
   • Make sure the provider matches your configured API key

2. "Failed to connect to API"

   • Check your internet connection
   • Verify your API key is valid for the selected provider
   • Check if there are any firewall restrictions

3. Keybindings not working

   • Ensure you've added the keybindings to your config.keys table as shown in the installation example
   • Check for conflicts with other keybindings in your config
   • Verify that local act = wezterm.action is defined in your config

4. Plugin not loading

   • Ensure the plugin directory structure is correct
   • Check WezTerm's configuration file syntax
   • Look for error messages in WezTerm's debug output

5. No commands generated

   • Try rephrasing your prompt to be more specific
   • Ensure your prompt describes a task that can be accomplished with bash commands

Security Notes

• Your API keys are stored in your WezTerm configuration and sent to the respective AI provider's servers (Anthropic or OpenAI)
• Prompts are processed by the selected AI provider's API to generate commands
• Consider the sensitivity of information you share with the AI
• Generated commands are not automatically executed - you choose when to run them
• Keep your API keys secure and don't commit them to version control

License

MIT