Skip to content

A lightweight building block for rapidly prototyping LLM-powered agents with customizable behaviors and pluggable tools.

License

Notifications You must be signed in to change notification settings

zhiyanliu/rapid-assistant

Repository files navigation

Rapid Assistant

Under development.

Installation

  • Install in editable mode (for your local development).
pip install -e . --config-settings editable_mode=strict
  • Install from the source code (before we publish the package to PyPI).
pip install git+https://github.com/rapid-assistant/rapid-assistant.git
  • Install from the PyPI (for production) - WIP
pip install rapid-assistant

Try it

# Install dependencies manually if you haven't followed the above instructions:
# cd rapid-assistant  # repo directory
# pip install -r requirements.txt

# Install dependencies for development, e.g. for yolo_test, repl_cli, websocket_server:
# pip install -r requirements-dev.txt

# Prepare .env file
cp .env.template .env
vi .env

# Config MCP servers as you needed
cp mcp_config.json.template mcp_config.json
vi mcp_config.json

# Run the assistant - single-turn mode
# python -m rapid_assistant.yolo_test --help
python -m rapid_assistant.yolo_test

# Run the assistant - human-in-the-loop/multi-turn mode
# python -m rapid_assistant.repl_cli --help
python -m rapid_assistant.repl_cli

# Run the assistant - websocket/UI mode
python -m rapid_assistant.websocket_server

About

A lightweight building block for rapidly prototyping LLM-powered agents with customizable behaviors and pluggable tools.

Topics

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages