Berget CLI

A command-line tool for interacting with Berget AI's infrastructure and AI models.

Installation

npm install -g berget
# or use directly with npx
npx berget --help

Authentication

Before you can use the CLI, you need to authenticate:

# Login with OAuth
npx berget auth login

# Create an API key
npx berget api-keys create --name "My CLI Key"

# Or use environment variable
export BERGET_API_KEY=sk_ber_your_api_key_here

Chat Command

Basic Usage

# Interactive chat session
npx berget chat run

# Use specific model
npx berget chat run openai/gpt-oss

# Send direct message
npx berget chat run openai/gpt-oss "Explain what Docker is"

# Use pipe for input
echo "What is Kubernetes?" | npx berget chat run openai/gpt-oss

Practical Use Cases

1. Git Commit Messages

# Generate commit message from git diff
git diff | npx berget chat run openai/gpt-oss "Create a conventional commit message for this diff. Reply with only the message:"

# Use as alias
alias gitcommit='git diff | npx berget chat run openai/gpt-oss "Generate a conventional commit message for this diff. Reply with only the commit message, nothing else:"'

2. Code Review and Explanations

# Explain code
cat src/main.js | npx berget chat run openai/gpt-oss "Explain what this code does:"

# Find bugs
cat problematic-file.py | npx berget chat run openai/gpt-oss "Analyze this code and find potential bugs:"

# Improvement suggestions
git diff | npx berget chat run openai/gpt-oss "Give suggestions for improvements to this code change:"

3. Documentation

# Generate README
ls -la | npx berget chat run openai/gpt-oss "Create a README.md for this project based on the file structure:"

# Comment code
cat uncommented-code.js | npx berget chat run openai/gpt-oss "Add JSDoc comments to this code:"

4. System Administration

# Analyze logs
tail -n 100 /var/log/nginx/error.log | npx berget chat run openai/gpt-oss "Analyze these error logs and suggest solutions:"

# Explain commands
npx berget chat run openai/gpt-oss "Explain what this bash command does: find . -name '*.js' -exec grep -l 'TODO' {} \;"

Useful Bash/Zsh Aliases

Add these to your ~/.bashrc, ~/.zshrc or similar:

# Git-related aliases
alias gai='git diff | npx berget chat run openai/gpt-oss "Generate a conventional commit message for this diff. Reply with only the commit message, nothing else:"'
alias gexplain='git log --oneline -10 | npx berget chat run openai/gpt-oss "Explain what these commits do:"'
alias gsec='~/bin/security-check'

# Code-related aliases
alias explain='npx berget chat run openai/gpt-oss "Explain this code:"'
alias review='npx berget chat run openai/gpt-oss "Review this code and give improvement suggestions:"'
alias debug='npx berget chat run openai/gpt-oss "Find and explain potential bugs in this code:"'

# Documentation aliases
alias docgen='npx berget chat run openai/gpt-oss "Generate documentation for this code:"'
alias readme='ls -la | npx berget chat run openai/gpt-oss "Create a README.md for this project:"'

# System aliases
alias loganalyze='npx berget chat run openai/gpt-oss "Analyze these logs and suggest solutions:"'
alias cmdexplain='npx berget chat run openai/gpt-oss "Explain this command:"'

# Quick AI assistant
alias ai='npx berget chat run openai/gpt-oss'
alias ask='npx berget chat run openai/gpt-oss'

Advanced Examples

See the examples/ folder for complete scripts:

• smart-commit.sh - Automatic generation of conventional commit messages
• ai-review.sh - AI-driven code review
• security-check.sh - Security review of commits

# Copy example scripts
cp examples/*.sh ~/bin/
chmod +x ~/bin/*.sh

# Use them
~/bin/smart-commit.sh
~/bin/ai-review.sh src/main.js
~/bin/security-check.sh

Environment Variables

# API key (recommended)
export BERGET_API_KEY=sk_ber_your_api_key_here

# Debug mode
export LOG_LEVEL=debug

# Custom API base URL (if using your own instance)
export API_BASE_URL=https://your-custom-api.example.com

Tips and Tricks

1. Use pipes: Combine with other Unix tools for powerful workflows
2. Short prompts: Be specific but concise in your prompts for best results
3. Streaming: Streaming is enabled by default for faster responses
4. Model selection: Experiment with different models for different tasks
5. Aliases: Create aliases for common use cases to save time

Command Reference

• auth login - Login to Berget
• auth logout - Logout from Berget
• auth whoami - Show current user information
• api-keys list - List API keys
• api-keys create - Create a new API key
• models list - List available AI models
• chat run - Start a chat session with an AI model
• chat list - List available chat models

For a complete list of commands, run:

npx berget --help

Troubleshooting

# Enable debug mode
npx berget --debug chat run openai/gpt-oss "test"

# Check authentication
npx berget auth whoami

# List available models
npx berget chat list

# Check API key status
npx berget api-keys list

Development

Setup

Clone the repository and install dependencies:

git clone https://github.com/berget-ai/cli.git
cd cli
npm install

Test Locally

Use the start script to test the CLI locally with the --local flag:

npm start -- <command> [options]

For example:

# Test login
npm start -- auth login

# Test whoami
npm start -- auth whoami

# Test with debug output
npm start -- auth whoami --debug

Contributing

Berget CLI is open source. Contributions are welcome!

• GitHub: berget-ai/cli
• Issues: Report bugs
• Documentation: docs.berget.ai