Hands-On AI Command Line Interface (CLI)¶
This document provides a reference for the command-line interfaces in Hands-On AI. The toolkit provides several CLI commands for different modules.
Main CLI (hands-on-ai)¶
The main hands-on-ai
command provides access to global functions and serves as an entry point to the module-specific commands.
Usage: hands-on-ai [OPTIONS] COMMAND [ARGS]...
Hands-On AI Toolkit
Options:
--install-completion Install completion for the current shell.
--show-completion Show completion for the current shell, to copy
it or customize the installation.
--help Show this message and exit.
Commands:
version Display version information
list List available modules
doctor Check environment and configuration
config View or edit configuration
Example Usage¶
# Check system configuration
hands-on-ai doctor
# View and edit configuration
hands-on-ai config
# List all available modules
hands-on-ai list
# Check version
hands-on-ai version
Chat Module CLI (chat)¶
The chat
module provides an interface for interacting with various bot personalities.
Usage: chat [OPTIONS] COMMAND [ARGS]...
Simple chatbot with personality
Options:
--install-completion Install completion for the current shell.
--show-completion Show completion for the current shell, to copy
it or customize the installation.
--help Show this message and exit.
Commands:
ask Send a single prompt to a bot
bots List available bots
doctor Run diagnostics
interactive Start interactive REPL
web Launch web interface
Example Usage¶
# List all available bot personalities
chat bots
# Ask a question with a specific bot personality
chat ask --personality pirate_bot "Tell me about the weather"
# Start an interactive chat session
chat interactive
# Launch the web interface for chat
chat web
RAG Module CLI (rag)¶
The rag
(Retrieval-Augmented Generation) module provides tools for document indexing and question answering.
Usage: rag [OPTIONS] COMMAND [ARGS]...
RAG - Retrieval-Augmented Generation
Options:
--install-completion Install completion for the current shell.
--show-completion Show completion for the current shell, to copy
it or customize the installation.
--help Show this message and exit.
Commands:
index Build a RAG index from files
ask Ask questions using indexed documents
interactive Run interactive RAG chat
web Launch web interface for RAG
Example Usage¶
# Index a directory of documents
rag index docs/
# Ask a question using the indexed documents
rag ask "What is the purpose of RAG?"
# Start an interactive RAG session
rag interactive
# Launch the web interface for RAG
rag web
Detailed Command Options¶
rag index¶
Usage: rag index [OPTIONS] INPUT_PATH
Build a RAG index from files
Arguments:
INPUT_PATH File or directory to index [required]
Options:
--output-file TEXT Output index file (default: ~/.hands-on-ai/index.npz)
--chunk-size INTEGER Words per chunk (default: from config)
--force / --no-force Overwrite existing index [default: no-force]
--help Show this message and exit.
Agent Module CLI (agent)¶
The agent
module provides tools for working with AI agents that can use tools to perform tasks.
Usage: agent [OPTIONS] COMMAND [ARGS]...
Agent - LLM Tool Use & Reasoning
Options:
--install-completion Install completion for the current shell.
--show-completion Show completion for the current shell, to copy
it or customize the installation.
--help Show this message and exit.
Commands:
ask Run an agent with a prompt
tools List available tools
interactive Start interactive agent console
web Launch web interface for agent
Example Usage¶
# List available agent tools
agent tools
# Run an agent with a specific prompt
agent ask "What is 25 * 32?"
# Start an interactive agent session
agent interactive
# Launch the web interface for the agent
agent web
Tips for Using the CLI¶
- All commands support the
--help
flag to display available options - Use tab completion (if installed) for easier navigation
- The web interfaces provide a graphical alternative to the command-line
- The interactive modes (REPL) allow for multi-turn conversations
- Configure settings in
~/.hands_on_ai/config.json
or with thehands-on-ai config
command