quick_start.md

Quick Start Guide

This guide provides the most direct path for a new user to get cli_engineer up and running.

Prerequisites

Before you begin, ensure you have the following installed:

• Rust and Cargo: cli_engineer is built with Rust. You can install the toolchain from the official website: https://www.rust-lang.org/tools/install

Installation

You can install cli_engineer directly from crates.io using Cargo:

cargo install cli_engineer

After installation, verify it's working by checking the help message:

cli_engineer --help

Configuration

cli_engineer requires configuration for an AI provider to function.

1. Create a Configuration File: Create a file named cli_engineer.toml in your project directory. You can start with this minimal example for OpenAI:

   # cli_engineer.toml
   
   [execution]
   max_iterations = 8
   
   [ui]
   colorful = true
   
   # Enable the provider you want to use.
   [ai_providers.openai]
   enabled = true
   model = "gpt-4o-mini" # A fast and affordable model
   temperature = 0.7
   cost_per_1m_input_tokens = 0.15
   cost_per_1m_output_tokens = 0.60
   max_tokens = 128000
   
   # Disable other providers
   [ai_providers.anthropic]
   enabled = false
   
   [ai_providers.openrouter]
   enabled = false
   
   [ai_providers.gemini]
   enabled = false
   
   [ai_providers.ollama]
   enabled = false

2. Set Your API Key: cli_engineer reads API keys from environment variables. For the configuration above, set your OpenAI key:

   export OPENAI_API_KEY="your-openai-api-key-here"

   For other providers, use the corresponding environment variable (ANTHROPIC_API_KEY, GEMINI_API_KEY, etc.).

First Run

Now you're ready to run your first task. Let's ask the agent to generate a simple Python script.

cli_engineer code "create a python script that implements the fizzbuzz algorithm and writes the output for numbers 1 to 100 to a file named fizzbuzz_output.txt"

The agent will now start, display its plan, and execute the steps to create the requested script. You will see the output in your terminal and find the generated files in the artifacts/ directory.