rabbit-ai-cli 0.0.12

A tiny, extensible AI CLI for terminal workflows

Rabbit 🐰

A tiny, extensible AI CLI for terminal workflows. Follow your curiosity with multi-provider AI chat, persistent sessions, and streaming responses.

Features

• Multi-Provider Support: OpenAI, Anthropic, and Ollama (local models)
• Persistent Sessions: Keep conversation history across interactions
• Streaming Output: Real-time token streaming with --stream
• System Prompts: Customize AI behavior with --system
• JSON Export: Machine-readable output format
• STDIN Support: Pipe input for automation
• Modular Design: Clean, testable architecture

Quick Start

Prerequisites

• Python 3.7+
• API keys for your chosen provider(s)

Installation

1. Clone the repository:

git clone <repository-url>
cd rabbit

2. Set up Python environment:

python3 -m venv .venv
source .venv/bin/activate  # On Windows: .venv\Scripts\activate
pip install -r requirements.txt  # If you have one, or install dependencies manually

3. Set up API keys:

Option A: Using environment file (recommended)

# Copy the example file and edit it with your keys
cp .env.example .env
# Edit .env with your favorite editor and add your API keys

Option B: Using environment variables

# For OpenAI
export OPENAI_API_KEY="your-openai-api-key"

# For Anthropic
export ANTHROPIC_API_KEY="your-anthropic-api-key"

# For Ollama (optional, defaults to localhost:11434)
export OLLAMA_HOST="http://localhost:11434"

Note: The CLI will first check for keys in a .env file, then fall back to environment variables.

Usage

Direct CLI Usage

# Simple question
python3 core/rabbit.py -q "What is Python?"

# With specific provider and model
python3 core/rabbit.py -q "Explain async/await" -p anthropic -m claude-3-sonnet-20240229

# Stream response
python3 core/rabbit.py -q "Write a Python function" --stream

# Use system prompt
python3 core/rabbit.py -q "Hello" --system "You are a helpful coding assistant"

# Start a persistent session
python3 core/rabbit.py -q "Let's discuss Python" -s python-chat

# Continue the session
python3 core/rabbit.py -q "Tell me about decorators" -s python-chat

# Show session history
python3 core/rabbit.py --show -s python-chat --limit 10

# JSON output for automation
python3 core/rabbit.py -q "Hello" --json

# Pipe input
echo "Explain this code: print('hello')" | python3 core/rabbit.py

Using the Wrapper Script

The bin/rabbit script provides a more convenient interface:

# Make it executable
chmod +x bin/rabbit

# Simple question
./bin/rabbit ask "What is Docker?"

# With additional flags
./bin/rabbit ask "Explain Kubernetes" -- -p openai -m gpt-4 --stream

# Show session history
./bin/rabbit show -s devops --limit 12

Configuration

Environment File

Rabbit supports loading API keys from a .env file for convenience and security. This is the recommended approach as it keeps your keys out of your shell history and makes it easy to manage different configurations.

Create a .env file in the project root:

# OpenAI Configuration
OPENAI_API_KEY=sk-your-actual-openai-key-here

# Anthropic Configuration  
ANTHROPIC_API_KEY=sk-ant-your-actual-anthropic-key-here

# Ollama Configuration (optional)
OLLAMA_HOST=http://localhost:11434

Priority Order:

1. .env file in the current directory
2. Environment variables
3. Default values (where applicable)

Security Note: Always add .env to your .gitignore to avoid committing API keys to version control.

Providers

Provider	Environment Variable	Default Model
OpenAI	OPENAI_API_KEY	gpt-4o-mini
Anthropic	ANTHROPIC_API_KEY	claude-3-5-sonnet-latest
Ollama	OLLAMA_HOST (optional)	llama3

Session Storage

Sessions are stored in ~/.config/aibot/sessions/ (or $XDG_CONFIG_HOME/aibot/sessions/).

Command Line Options

-q, --query QUERY       Your question/prompt
-s, --session NAME      Session name for persistent conversation
-p, --provider PROVIDER Provider: openai, anthropic, ollama
-m, --model MODEL       Model name (provider-specific)
--system PROMPT         System prompt to set behavior
--stream                Stream tokens as they arrive
--json                  Output JSON envelope
--show                  Show session history
--limit N               Number of messages to show (with --show)

Development

Project Structure

rabbit/
├── bin/
│   └── rabbit              # Convenience wrapper script
├── core/
│   ├── __init__.py         # Package marker
│   ├── rabbit.py           # Main CLI entry point
│   ├── config.py           # Configuration paths
│   ├── io_utils.py         # Terminal I/O utilities
│   ├── messages.py         # Message construction
│   ├── sessions.py         # Session persistence
│   └── providers/
│       └── __init__.py     # Provider adapters
└── tests/
    └── test_rabbit.py      # Unit tests

Running Tests

# Install test dependencies
pip install pytest

# Run all tests
python -m pytest tests/ -v

# Run specific test file
python -m pytest tests/test_rabbit.py -v

# Run with coverage (if pytest-cov installed)
python -m pytest tests/ --cov=core --cov-report=html

Adding New Providers

1. Add your provider function to core/providers/__init__.py:

def call_your_provider(messages: List[Message], model: str, stream: bool) -> str:
    # Your implementation here
    pass

2. Register it in the provider dictionaries:

PROVIDER_DEFAULTS["your_provider"] = "default-model"
PROVIDERS["your_provider"] = call_your_provider

3. Add tests in tests/test_rabbit.py

Examples

Basic Q&A

python3 core/rabbit.py -q "What's the difference between list and tuple in Python?"

Code Review Session

# Start a code review session
python3 core/rabbit.py -q "I need help reviewing some Python code" -s code-review

# Continue in the same session
python3 core/rabbit.py -q "Here's the function: def calculate_total(items): ..." -s code-review

# Check what we discussed
python3 core/rabbit.py --show -s code-review

Automated Processing

# Process multiple files
for file in *.py; do
    echo "# Reviewing $file" 
    cat "$file" | python3 core/rabbit.py --system "You are a code reviewer" --json
done

Local AI with Ollama

# Make sure Ollama is running locally with a model
ollama pull llama3

# Use local model
python3 core/rabbit.py -q "Hello" -p ollama -m llama3

Installation from PyPI

Once published, you can install rabbit directly from PyPI:

pip install rabbit-ai-cli

Then use it directly:

rabbit -q "Hello world" -p openai

Development & Publishing

Local Development

1. Clone and set up the development environment:

git clone https://github.com/zeecaniago/rabbit.git
cd rabbit
python3 -m venv .venv
source .venv/bin/activate
pip install -r requirements.txt

2. Test your changes:

python -m pytest tests/

3. Test package building:

./test_build.sh

Publishing Process

This project uses automated publishing to PyPI:

1. Automatic Tagging: Every commit to main automatically creates a new version tag (managed by .github/workflows/auto-tag.yml)

2. Automatic Publishing: When a new tag is created, the publish workflow (.github/workflows/publish.yml) automatically:

   • Builds the package
   • Updates the version to match the git tag
   • Runs tests
   • Publishes to PyPI
   • Creates a GitHub release

Setting up PyPI Publishing

To enable automatic publishing, you need to:

1. Create a PyPI account at https://pypi.org

2. Set up Trusted Publishing (recommended):

   • Go to https://pypi.org/manage/account/publishing/
   • Add a new "pending publisher" with:
     • PyPI project name: rabbit-ai-cli
     • Owner: zeecaniago
     • Repository name: rabbit
     • Workflow name: publish.yml
     • Environment name: (leave empty)

3. Alternative: Use API tokens:

   • Generate an API token at https://pypi.org/manage/account/token/
   • Add it as PYPI_API_TOKEN in your repository secrets

The workflow will automatically handle version management based on your git tags.

Contributing

1. Fork the repository
2. Create a feature branch
3. Add tests for new functionality
4. Ensure all tests pass
5. Submit a pull request

License

[Your chosen license]

Troubleshooting

Common Issues

"Import could not be resolved" errors: These are lint warnings for optional dependencies (openai, anthropic, requests). The code handles missing imports gracefully.

"Command not found: pytest": Install pytest with pip install pytest

API key errors: Make sure your API keys are set in environment variables and are valid.

Ollama connection errors: Ensure Ollama is running locally and the model is available.

Getting Help

• Check existing issues in the repository
• Run with -h for help: python3 core/rabbit.py -h
• Use --show to debug session issues
• Check API key configuration with your provider's documentation