Lab1: Setting Up Google Gemini API

Introduction

In this lab, you will learn how to set up Google Gemini API using classroom credits. Gemini is Google’s multimodal LLM that can understand text, images, and more, making it ideal for robotics applications.

Prerequisites

  • Temple University email address (for credits)
  • Personal Gmail account
  • Mini Pupper or development computer

Part 1: Request Google Cloud Credits

Step 1: Request Credits from Canvas

  1. Go to Canvas and request your Google credit using your temple.edu email address
Request coupon from Canvas

Step 2: Check Your Email

  1. Check your Temple email for the coupon message
  2. Save the coupon code
  3. Copy the redeem link
Coupon email message

Part 2: Create Google Cloud Project

Step 1: Login with Personal Gmail

  1. Open a new Chrome tab
  2. Login with your personal Gmail account (not Temple email)

Step 2: Create New Project

Use the redeem link to create a new project:

Create project step 1
Create project step 2

Step 3: Redeem Credits

  1. Use the redeem link in a new tab
  2. Select your project
  3. Paste the coupon code
Setup coupon

Part 3: Create Service Account

Go to Google Credentials and make sure you select your personal Gmail account.

Step 1: Create Credential

Create credential

Step 2: Service Account

Service account

Step 3: Create Keys

Create keys

Step 4: Add Key

Add key

Step 5: Download JSON

Download JSON

Important:

  • SAVE this JSON file and keep it private!
  • Rename it as minipupper_creds.json
  • Upload the file to the home directory on the robot

Part 4: Enable Required APIs

You need to enable the following APIs:

  • Vertex AI API
  • Cloud Speech-to-Text API
  • Cloud Text-to-Speech API

Step 1: Go to API/Services

API services

Step 2: Enable APIs

Enable APIs

Step 3: Search for API

Search API

Step 4: Enable the API

Enable API

Step 5: Verify Status

API status

Repeat for all three APIs (Vertex AI, Speech-to-Text, Text-to-Speech).

Part 5: Check Your Credits

Go to Google Cloud Console > Billing

Step 1: Billing API

Billing API

Step 2: Manage Account

Manage account

Step 3: Select Account

Select account

Step 4: View Cost

View cost

Part 6: Setup Credentials on Mini Pupper

Step 1: Create Credential Directory 

mkdir -p ~/.gemini

Step 2: Upload and Move Credential File

Upload minipupper_creds.json to the robot, then:

cp ~/minipupper_creds.json ~/.gemini/creds.json

Step 3: Set Environment Variable

Add to your ~/.bashrc:

nano ~/.bashrc

# Add at the end:
export GOOGLE_APPLICATION_CREDENTIALS=/home/ubuntu/.gemini/creds.json

# Save and apply
source ~/.bashrc

Part 7: Install Gemini App

Step 1: Clone the Repository 

cd ~
git clone https://github.com/lbaitemple/apps-md-robots
cd apps-md-robots
pip install -r requirements.txt

Step 2: Setup Environment File 

cp env.sample .env

Step 3: Run the AI App 

python ai_app/ai_app.py

Step 4: Reset LCD (if needed) 

sudo systemctl restart robot

Part 8: Language Support

Google Cloud Text-to-Speech supports many languages:

Language support

See full list at: Cloud Text-to-Speech Voices

Summary

In this lab, you learned:

  • How to request and redeem Google Cloud classroom credits
  • How to create a Google Cloud project
  • How to create a service account and download credentials
  • How to enable required APIs (Vertex AI, Speech-to-Text, Text-to-Speech)
  • How to configure credentials on Mini Pupper
  • How to run the Gemini AI app
  • How to test Gemini Vision with food analysis

Reference