Skip to main content

​
Overview

The CLI provides an interactive REPL (Read-Eval-Print Loop) interface for natural language interactions with your Google Workspace services. It supports both Legacy and DSPy prompt modes.

​
Starting the CLI

1

Run the main script

python main.py
You’ll see the welcome message:
AgenticPal - Your AI Personal Assistant
Type 'exit', 'quit', or 'q' to exit
Mode: Legacy
--------------------------------------------------

You:
2

Enter your request

Type your request in natural language and press Enter.

​
Configuration Options

Before running, you can configure the prompt mode in main.py: 
# Use traditional LangChain prompts - more stable
use_legacy()

​
Example Interactions

​
Calendar Management

You: Add a meeting with John next Tuesday at 2pm

Assistant: I've created a new event "Meeting with John" on Tuesday, March 11, 2026 at 2:00 PM.

You: What's on my calendar this week?

Assistant: You have 3 events this week:
1. Team Standup - Monday at 9:00 AM
2. Meeting with John - Tuesday at 2:00 PM
3. Project Review - Friday at 3:00 PM

You: Move my meeting with John to 3pm

Assistant: I've updated "Meeting with John" to 3:00 PM on Tuesday, March 11, 2026.

​
Email Operations

You: Summarize my emails from this week

Assistant: You have 12 emails from this week:
- 5 from your team regarding the project launch
- 3 newsletter subscriptions
- 2 customer support inquiries
- 2 meeting invitations

You: Find emails from Sarah about the budget

Assistant: I found 2 emails from Sarah about the budget:
1. "Q1 Budget Review" - March 5, 2026
2. "Budget Approval Request" - March 7, 2026

You: Send an email to [email protected] with subject "Status Update" and say the project is on track

Assistant: I've sent the email to [email protected] with subject "Status Update".

​
Task Management

You: Create a task: buy groceries

Assistant: I've created a new task "buy groceries" in your default task list.

You: What tasks do I have?

Assistant: You have 4 incomplete tasks:
1. Buy groceries
2. Review pull requests
3. Prepare presentation
4. Call dentist

You: Mark "buy groceries" as done

Assistant: I've marked "buy groceries" as completed.

​
Multi-turn Conversations

The agent maintains conversation context and can ask clarifying questions: 
You: Schedule a meeting with the team

Assistant: I'd be happy to help schedule a meeting. Could you provide:
1. What date and time would you like?
2. How long should the meeting be?
3. What's the meeting topic?

You: Tomorrow at 10am for 1 hour about the project launch

Assistant: I've created a 1-hour meeting "Project Launch" for tomorrow at 10:00 AM with the team.

​
Special Commands

Help

Type help to see available tools and their descriptions
You: help

Exit

Type exit, quit, or q to close the CLI
You: exit
Goodbye!

​
Error Handling

The CLI gracefully handles errors and continues running: 
You: Schedule a meeting for tomorrow at invalid-time

Error: Could not parse the time "invalid-time". Please use a format like "2pm" or "14:00".

You:
Press Ctrl+C at any time to gracefully exit the CLI.

​
Configuration

The graph runner is initialized with these default settings: 
graph, tools_registry = create_graph_runner(
    calendar_service=calendar_service,
    gmail_service=gmail_service,
    tasks_service=tasks_service,
    model_name="qwen-plus-2025-12-01",
    default_timezone="UTC"
)
You can customize the model_name and default_timezone parameters in main.py to match your preferences.

Build docs developers (and LLMs) love

Get started for freeTalk to us