basic-agent-chat-loop 0.1.3

Feature-rich interactive CLI for AI agents with token tracking, prompt templates, aliases, and configuration

Basic Agent Chat Loop

A feature-rich, interactive CLI for AI agents with token tracking, prompt templates, agent aliases, and extensive configuration options.

Features

• 🏷️ Agent Aliases - Save agents as short names (chat_loop pete instead of full paths)
• 📜 Command History - Navigate previous queries with ↑↓ arrows (persisted to ~/.chat_history)
• ✍️ Multi-line Input - Type \\ to enter multi-line mode for code blocks
• 💰 Token Tracking - Track tokens and costs per query and session
• 📝 Prompt Templates - Reusable prompts from ~/.prompts/
• ⚙️ Configuration - YAML-based config with per-agent overrides
• 📊 Status Bar - Real-time metrics (queries, tokens, duration)
• 📈 Session Summary - Full statistics displayed on exit
• 🎨 Rich Formatting - Enhanced markdown rendering with syntax highlighting
• 🔄 Error Recovery - Automatic retry logic with exponential backoff
• 🔍 Agent Metadata - Display model, tools, and capabilities

Installation

Quick Install (Recommended)

pip install basic-agent-chat-loop

That's it! The package will automatically create:

• ~/.chatrc - Configuration file with recommended defaults
• ~/.prompts/ - Sample prompt templates (on first use)

Platform-Specific Options

Windows (for command history support):

pip install basic-agent-chat-loop[windows]

AWS Bedrock integration:

pip install basic-agent-chat-loop[bedrock]

From Source

For development or the latest features:

git clone https://github.com/Open-Agent-Tools/Basic-Agent-Chat-Loop.git
cd Basic-Agent-Chat-Loop
pip install -e ".[dev]"

See docs/INSTALL.md for detailed installation instructions and troubleshooting.

Quick Start

Basic Usage

# Run with agent path
chat_loop path/to/your/agent.py

# Or use an alias (after saving)
chat_loop myagent

Agent Aliases

Save frequently used agents for quick access:

# Save an agent as an alias
chat_loop --save-alias myagent path/to/agent.py

# Use the alias from anywhere
chat_loop myagent

# List all saved aliases
chat_loop --list-aliases

# Remove an alias
chat_loop --remove-alias myagent

Example with real agents:

# Save your agents
chat_loop --save-alias pete ~/agents/product_manager/agent.py
chat_loop --save-alias dev ~/agents/senior_developer/agent.py

# Use them from anywhere
cd ~/projects/my-app
chat_loop dev  # Get coding help
chat_loop pete  # Get product feedback

Aliases are stored in ~/.chat_aliases and work from any directory.

Prompt Templates

The package automatically creates sample templates in ~/.prompts/ on first use:

• explain.md - Explain code in detail
• review.md - Code review with best practices
• debug.md - Help debugging issues
• optimize.md - Performance optimization suggestions
• test.md - Generate test cases
• document.md - Add documentation

Use templates in chat:

chat_loop myagent
You: /review src/app.py
You: /explain utils.py
You: /test my_function

Create custom templates:

# Create your own template
cat > ~/.prompts/security.md <<'EOF'
# Security Review

Please review this code for security vulnerabilities:

{input}

Focus on:
- Input validation
- Authentication/authorization
- Data sanitization
- Common security patterns
EOF

# Use it in chat
You: /security auth.py

Configuration

A configuration file (~/.chatrc) is automatically created on first use with recommended defaults. You can customize it to your preferences:

features:
  show_tokens: true           # Display token counts
  show_metadata: true         # Show agent model/tools info
  rich_enabled: true          # Enhanced formatting

ui:
  show_status_bar: true       # Top status bar
  show_duration: true         # Query duration

behavior:
  max_retries: 3              # Retry attempts on failure
  timeout: 120.0              # Request timeout (seconds)

# Per-agent overrides
agents:
  'Product Pete':
    features:
      show_tokens: false

See CONFIG.md for full configuration options.

Commands

Command	Description
help	Show help message
info	Show agent details (model, tools)
templates	List available prompt templates
/name	Use prompt template from ~/.prompts/name.md
clear	Clear screen and reset agent session
exit, quit	Exit chat (shows session summary)

Multi-line Input

Press \\ to enter multi-line mode:

You: \\
... def factorial(n):
...     if n <= 1:
...         return 1
...     return n * factorial(n - 1)
...
[Press Enter on empty line to submit]

Token Tracking

During Chat

When show_tokens: true in config:

------------------------------------------------------------
Time: 6.3s │ 1 cycle │ Tokens: 4.6K (in: 4.4K, out: 237) │ Cost: $0.017

Session Summary

Always shown on exit:

============================================================
Session Summary
------------------------------------------------------------
  Duration: 12m 34s
  Queries: 15
  Tokens: 67.8K (in: 45.2K, out: 22.6K)
  Total Cost: $0.475
============================================================

Programmatic Usage

from basic_agent_chat_loop import ChatLoop

# Create chat interface
chat = ChatLoop(
    agent=your_agent,
    name="My Agent",
    description="Agent description",
    config_path=Path("~/.chatrc")  # Optional
)

# Run interactive loop
chat.run()

Requirements

Core Dependencies

• Python 3.8+
• pyyaml>=6.0.1 - Configuration file parsing
• rich>=13.7.0 - Enhanced terminal rendering
• python-dotenv>=1.0.0 - Environment variable management

Optional Dependencies

• pyreadline3>=3.4.1 - Command history on Windows (install with [windows])
• anthropic-bedrock>=0.8.0 - AWS Bedrock integration (install with [bedrock])

Built-in Features

• readline (built-in on Unix) - Command history on macOS/Linux

Platform Support

• ✅ macOS - Full support with native readline
• ✅ Linux - Full support with native readline
• ✅ Windows - Full support with pip install basic-agent-chat-loop[windows]

Architecture

src/basic_agent_chat_loop/
├── chat_loop.py          # Main orchestration
├── chat_config.py        # Configuration management
├── cli.py                # CLI entry point
├── components/           # Modular components
│   ├── ui_components.py      # Colors, StatusBar
│   ├── token_tracker.py      # Token/cost tracking
│   ├── template_manager.py   # Prompt templates
│   ├── display_manager.py    # Display formatting
│   ├── agent_loader.py       # Agent loading
│   └── alias_manager.py      # Alias management
docs/
├── ALIASES.md            # Alias system guide
├── CONFIG.md             # Configuration reference
├── INSTALL.md            # Installation instructions
└── Chat_TODO.md          # Roadmap and future features

Documentation

• docs/ALIASES.md - Agent alias system guide
• docs/CONFIG.md - Configuration reference
• docs/INSTALL.md - Installation instructions
• docs/Chat_TODO.md - Roadmap and future features

Development

Running Tests

# Install dev dependencies
pip install -e ".[dev]"

# Run tests
pytest

Code Quality

# Format code
black src/ tests/

# Lint
ruff check src/ tests/

Contributing

Contributions are welcome! Please feel free to submit a Pull Request.

License

MIT License - see LICENSE file for details.

Changelog

See CHANGELOG.md for detailed version history.

Latest Release: v0.1.0 (2025-10-09)

Initial public release with:

• 🏷️ Agent alias system
• 📝 Prompt templates with auto-setup
• 💰 Token tracking and cost estimation
• ⚙️ YAML configuration with auto-setup
• 📊 Status bar and session summaries
• 🎨 Rich markdown rendering
• 🔄 Automatic error recovery
• 📜 Persistent command history
• ✅ 61% test coverage (158 tests)

Troubleshooting

See docs/TROUBLESHOOTING.md for common issues and solutions.

Quick fixes:

• Package not found: Run pip install --upgrade basic-agent-chat-loop
• Command not found: Ensure pip's bin directory is in your PATH
• Import errors: Try reinstalling with pip install --force-reinstall basic-agent-chat-loop

Support

• 🐛 Bug Reports: GitHub Issues
• 💡 Feature Requests: GitHub Issues
• 📖 Documentation: docs/
• 💬 Discussions: GitHub Discussions