Lite Package CLI

CLI commands for the lightweight agent package

Lite Package CLI

Commands for using the lightweight praisonaiagents.lite package from the command line.

Commands

Show Package Info

Display information about the lite package:

praisonai lite info

Output:

praisonaiagents.lite - Lightweight Agent Package
==================================================

Classes: LiteAgent, LiteTask, LiteToolResult
Decorators: @tool
Helpers: create_openai_llm_fn, create_anthropic_llm_fn

Features:
  • BYO-LLM (Bring Your Own LLM)
  • Thread-safe chat history
  • Tool execution
  • No litellm dependency
  • Minimal memory footprint

Show Example Code

Display example usage code:

praisonai lite example

Output:

# Example: Using praisonaiagents.lite with custom LLM

from praisonaiagents.lite import LiteAgent, tool

# Define a custom LLM function
def my_llm(messages):
    # Your custom LLM implementation
    pass

# Or use the built-in OpenAI adapter
from praisonaiagents.lite import create_openai_llm_fn
llm_fn = create_openai_llm_fn(model="gpt-4o-mini")

# Create a lite agent
agent = LiteAgent(
    name="MyAgent",
    llm_fn=llm_fn,
    instructions="You are a helpful assistant."
)

# Chat with the agent
response = agent.chat("Hello!")
print(response)

Run Lite Agent

Run a lite agent with a prompt:

# Using OpenAI (default)
praisonai lite run "Hello, how are you?"

# Specify model
praisonai lite run "Hello" --model gpt-4o-mini

# Use Anthropic
praisonai lite run "Hello" --provider anthropic --model claude-3-5-sonnet-20241022

Options:

Option	Description	Default
--model	Model name to use	gpt-4o-mini
--provider	LLM provider (openai, anthropic)	openai

Environment Variables

Required environment variables based on provider:

# For OpenAI
export OPENAI_API_KEY="your-key"

# For Anthropic
export ANTHROPIC_API_KEY="your-key"

Examples

Quick Chat

# Simple chat with OpenAI
export OPENAI_API_KEY="your-key"
praisonai lite run "What is 2+2?"

Using Different Models

# GPT-4o
praisonai lite run "Explain quantum computing" --model gpt-4o

# Claude
praisonai lite run "Write a haiku" --provider anthropic --model claude-3-5-sonnet-20241022

Check Availability

# Verify lite package is available
praisonai lite info

Comparison with Full CLI

Feature	praisonai	praisonai lite
Multi-provider	Yes (via litellm)	Manual only
Memory usage	~93MB	~5MB
Startup time	~800ms	~18ms
Dependencies	Many	Minimal

Related

• Lite Package (Code)
• Lazy Imports
• Performance CLI