🖥️ LLM-Term

A Rust-based CLI tool that generates and executes terminal commands using OpenAI's language models or local Ollama models.

Features

Configurable model and token limit (gpt-4o-mini, gpt-4o, or Ollama)
Generate and execute terminal commands based on user prompts
Works on both PowerShell and Unix-like shells (Automatically detected)

Demo

Installation

Download the binary from the Releases page
Set PATH to the binary
- MacOS/Linux:
```
export PATH="$PATH:/path/to/llm-term"
```
- To set it permanently, add export PATH="$PATH:/path/to/llm-term" to your shell configuration file (e.g., .bashrc, .zshrc)
- Windows:
```
set PATH="%PATH%;C:\path\to\llm-term"
```
- To set it permanently, add set PATH="%PATH%;C:\path\to\llm-term" to your shell configuration file (e.g., $PROFILE)

Development

Clone the repository
Build the project using Cargo: cargo build --release
The executable will be available in the target/release directory

Usage

Set your OpenAI API key (if using OpenAI models):
- MacOS/Linux:
```
export OPENAI_API_KEY="sk-..."
```
- Windows:
```
set OPENAI_API_KEY="sk-..."
```
If using Ollama, make sure it's running locally on the default port (11434)
Run the application with a prompt:
```
./llm-term "your prompt here"
```
The app will generate a command based on your prompt and ask for confirmation before execution.

Configuration

A config.json file will be created in the same directory as the binary on first run. You can modify this file to change the default model and token limit.

Options

-c, --config <FILE>: Specify a custom config file path

Supported Models

OpenAI GPT-4 (gpt-4o)
OpenAI GPT-4 Mini (gpt-4o-mini)
Ollama (local models, default: llama3.1)

dh1011/llm-term