A Python package for conversational AI interactions, enabling seamless chat-based interactions with AI models and RAG (Retrieval-Augmented Generation) capabilities.

• Quick chat interactions with AI models
• Support for RAG (Retrieval-Augmented Generation) with directory-based context
• Flexible conversation management and storage
• CLI interface for easy interactions
• Configurable AI model settings and API endpoints
• File-based chat history management

This project uses Poetry for dependency management. To install:

# Clone the repository
git clone https://github.com/yourusername/talkie.git
cd talkie

# Install with Poetry
poetry install

• Python 3.12 or higher
• OpenAI API key

1. Create a .env file in the project root:

OPENAI_API_KEY=your_api_key_here

2. The package uses default configurations that can be customized through the CLI or programmatic interface.

Talkie uses a hierarchical configuration system, where settings are applied in the following order (highest priority first):

1. Command-line Arguments

   • Arguments passed directly to commands (e.g., --api-key, --rag-dir)
   • Override all other settings

2. Chat File Frontmatter

   • Settings in individual .md chat files
   • Applies only to that specific chat
   • Overrides global config and defaults

3. Local Config File

   • config.talkie.yml in the current directory
   • Project-specific settings

4. Global Config File

   • ~/.talkie/config.talkie.yml
   • User-wide settings
   • Created automatically with defaults if not present

5. Default Values

   • Built-in defaults:

     system_prompt: "You are a helpful AI assistant..."
     model: "gpt-4"
     api_endpoint: "https://api.openai.com/v1/chat/completions"
     rag_directory: ""

You can customize the global configuration by editing ~/.talkie/config.talkie.yml, or create a local config.talkie.yml in your project directory for project-specific settings.

Talkie uses environment variables for sensitive information like API keys. The API key is looked up in the following order:

1. Command-line Argument

   • Passed via --api-key flag
   • Gets stored in ~/.talkie/.env for future use
   • Highest priority

2. Local .env File

   • .env file in your current directory
   • Good for project-specific keys

   OPENAI_API_KEY=sk-your-key-here
   

3. Environment Variable

   • Set in your shell session

   # Mac/Linux
   export OPENAI_API_KEY=sk-your-key-here
   
   # Windows (PowerShell)
   $env:OPENAI_API_KEY='sk-your-key-here'

4. Global .env File

   • ~/.talkie/.env file
   • Created automatically when using --api-key
   • Good for user-wide settings

If no API key is found, Talkie will display instructions for setting up your API key using any of these methods.

from talkie.chat.quick import quick_chat

# Simple question
response = await quick_chat("What is the capital of France?")

# With RAG context from a directory
response = await quick_chat(
    question="What are the key points in the documentation?",
    rag_directory="./docs"
)

# Quick chat - Get immediate responses without creating chat files
talkie quick "What is the weather like today?"
talkie quick "Explain this code" --rag-dir ./src  # Include codebase context
talkie quick "Help me debug" --output debug.md    # Save conversation to file
talkie quick "Write a poem" --system "You are a poet"  # Custom system prompt

# Create a new chat file - Initialize a markdown file for ongoing conversations
talkie create python-help                    # Create python-help.md in current dir
talkie create code-review --dir ~/chats/work # Create in specific directory

# Ask questions in existing chats - Continue conversations in markdown files
talkie ask python-help.md                    # Use default API key from env/config
talkie ask code-review.md --api-key sk-xxx   # Use specific API key

# Additional Commands
talkie hey "Quick greeting"                  # Send a quick greeting
talkie bye                                   # End the conversation

Each chat file (.md) is initialized with frontmatter containing:

• Title and timestamps
• System prompt and model settings
• API endpoint configuration
• RAG directory settings
• Tags and summary fields

You can customize the chat behavior by editing the frontmatter section at the top of any chat file:

---
title: python-help
system: You are a helpful Python programming assistant  # Change AI personality
model: gpt-4                                          # Switch between models
api_endpoint: https://api.openai.com/v1/chat/completions
rag_directory: ./src                                  # Add code context
created_at: 2024-01-20T10:30:00
updated_at: 2024-01-20T10:45:00
tags: [python, debugging]                            # Add organizational tags
summary: Debugging session for API implementation     # Add chat summary
---

For development work:

# Install dev dependencies
poetry install --with dev

# Run tests
pytest

# Format code
black .

• chat/: Core chat functionality and AI interactions
• cli/: Command-line interface implementations
• fsutils/: File system utilities
• rag/: Retrieval-Augmented Generation functionality

MIT License

1. Fork the repository
2. Create your feature branch (git checkout -b feature/amazing-feature)
3. Commit your changes (git commit -m 'Add some amazing feature')
4. Push to the branch (git push origin feature/amazing-feature)
5. Open a Pull Request