CodeAI CLI

A high-performance Go CLI tool that integrates local LLMs (via Ollama API) for code analysis, context-aware refactoring suggestions, and sophisticated prompt engineering within developer workflows.

Features

• Code Analysis: Analyze code for quality, potential issues, and improvement opportunities
• Refactoring Suggestions: Get AI-powered suggestions to improve your code
• Documentation Generation: Automatically generate comprehensive documentation for your code
• Code Explanation: Get detailed explanations of complex code snippets
• Test Generation: Generate comprehensive tests for your code
• Performance Optimization: Get suggestions to optimize code performance
• Security Analysis: Identify potential security vulnerabilities in your code
• Code Comparison: Compare two code files and understand semantic differences
• Docstring Generation: Generate documentation for specific functions or methods
• Project Analysis: Analyze entire projects or directories of code files
• Local LLM Integration: Uses Ollama to run models locally for privacy and performance

Prerequisites

• Go 1.16 or higher
• Ollama installed and running locally

Installation

# Clone the repository
git clone https://github.com/yourusername/cli-tool-go.git
cd cli-tool-go

# Build the binary
go build -o codeai

# Move to a directory in your PATH (optional)
mv codeai /usr/local/bin/

Setup

1. Install Ollama from ollama.ai
2. Pull a model (e.g., ollama pull llama2 or ollama pull gemma:7b)
3. Ensure Ollama is running (ollama serve in a terminal window)
4. Configure default settings (optional):

   codeai config set ollama.host http://localhost:11434
   codeai config set ollama.model llama2

Usage

List Available Models

codeai models

Analyze Code

codeai analyze path/to/file.go --context "This is a web server handler" --stream

Get Refactoring Suggestions

codeai refactor path/to/file.go --requirements "Improve error handling" --save

Generate Documentation

codeai document path/to/file.go --output documentation.md

Explain Code

codeai explain path/to/file.go --output explanation.md

Generate Tests

codeai gentest path/to/file.go --requirements "Include tests for edge cases" --save

Optimize Performance

codeai optimize path/to/file.go --focus "Time complexity and memory usage"

Security Analysis

codeai security path/to/file.go --language "Go"

Compare Code Files

codeai compare original.go updated.go --output comparison.md

Generate Function Docstring

# Generate docstring for a specific function
codeai docstring path/to/file.go functionName --clipboard

# Generate based on line numbers
codeai docstring path/to/file.go --line-start 10 --line-end 30

Analyze Entire Project

codeai analyze-project ./myproject --context "This is a REST API" --extensions go,md

Configure Settings

# View configuration
codeai config

# Set a configuration value
codeai config set ollama.model gemma3:12b

# Get a configuration value
codeai config get ollama.host

Command Options

Most commands support the following options:

• --model or -m: Specify which LLM model to use
• --temperature or -t: Control randomness (0.0-1.0)
• --stream or -S: Stream the response in real-time
• --output or -o: Save results to a file
• --context or -c: Provide additional context

Model Support

CodeAI works with any model available in Ollama, including:

• Llama 2/3
• Gemma/Gemma 3
• Mistral/Mixtral
• CodeLlama
• And many more

Development

Project Structure

codeai/
├── cmd/               # CLI commands
├── pkg/
│   ├── ollama/        # Ollama API client
│   ├── template/      # Prompt templates
│   └── clipboard/     # Clipboard utilities
├── examples/          # Example code files
├── main.go            # Entry point
└── README.md          # Documentation

Adding a New Command

1. Create a new file in the cmd directory
2. Define your command structure and logic
3. Add command to root.go

Example:

var newCommand = &cobra.Command{
    Use:   "newcommand [file]",
    Short: "Short description",
    Long:  `Longer description...`,
    Args:  cobra.ExactArgs(1),
    RunE: func(cmd *cobra.Command, args []string) error {
        // Command logic here
        return nil
    },
}

func init() {
    rootCmd.AddCommand(newCommand)
    // Add flags
}

Error Handling

The tool includes sophisticated error handling for common scenarios:

• Ollama not running or unavailable
• Invalid file paths
• Model loading issues
• Network connectivity problems

If Ollama is not running, the tool will provide helpful instructions to start it.

License

MIT License