Setup - Google Colab & the Gemini API

Generative AI for Managers - TUHH Institute of Entrepreneurship · part of my Technology Management MBA · study notes for revision.

---

Every lab in this course runs in Google Colab (a free, browser-based notebook - no installs on my laptop) and calls the Gemini API (Google’s LLM service). This page is the one-time setup so the notebooks in Days 1-5 just work. It takes about 15 minutes.

Note

What each piece is:

• Google Colab - a website that runs Python notebooks in the cloud. A notebook is a document mixing text and runnable code cells.
• Gemini API - the interface my code uses to send prompts to Google’s model and get answers back.
• API key - a secret password that identifies my account to the API. Anyone who has it can spend against my quota, so it must be protected.

What I need

A laptop + modern browser (Chrome)A Google account

Step 1 - Open Google Colab

1. Go to colab.research.google.com and sign in with my Google account.

2. Click New Notebook.

3. In the first cell, type this and run it (Shift+Enter):

   print("Colab is working!")

   If it prints the message, Colab is ready.

Step 2 - Create a Gemini API key

1. Open Google AI Studio and sign in.

2. Find Get API key (or API keys) in the interface and create a new API key.

3. Save it somewhere safe - a password manager is ideal. Treat it exactly like a password.

Danger

Never expose your API key. Do not paste it into a shared notebook, a screenshot, a chat, or a document - and never commit it to GitHub. If a key ever leaks (even inside a file you shared), assume it’s compromised and delete it in AI Studio and create a new one. The safe pattern below means the key never appears in the notebook’s text at all.

Note

Free-tier availability and rate limits depend on your region and account and can change. Keep usage small and follow the in-class guidance.

Step 3 - Store the key safely in Colab (Secrets)

The right way to give a notebook the key without writing it into the code:

1. In Colab, click the key icon (🔑 Secrets) in the left sidebar.

2. Add a new secret - Name: GEMINI_API_KEY, Value: paste your key.

3. Toggle Notebook access on for that secret.

4. In a code cell, load it into an environment variable:

   from google.colab import userdata
   import os
   
   os.environ["GEMINI_API_KEY"] = userdata.get("GEMINI_API_KEY")
   print("Key loaded:", bool(os.environ.get("GEMINI_API_KEY")))

   If it prints Key loaded: True, you’re set.

Tip

Fallback (only in your own private notebook): if Secrets aren’t available, paste the key once per session without it showing on screen:

import os, getpass
os.environ["GEMINI_API_KEY"] = getpass.getpass("Paste your GEMINI_API_KEY: ")
print("Key loaded:", bool(os.environ.get("GEMINI_API_KEY")))

getpass hides the input as you type - still, never save the key into the notebook itself.

Step 4 - Install the SDK and make a first call

1. Install the Gemini SDK (the ! runs a shell command inside Colab):

   !pip -q install -U google-genai

2. Make a first call. genai.Client() automatically reads GEMINI_API_KEY from the environment:

   import os
   from google import genai
   
   client = genai.Client()  # reads GEMINI_API_KEY from the environment
   resp = client.models.generate_content(
       model="gemini-2.5-flash-lite",
       contents="Say hello in one short sentence."
   )
   print(resp.text)

   If you see a greeting, the whole setup works.

Note

Model name: the labs use gemini-2.5-flash-lite - a small, fast, cheap model that’s plenty for learning. You’ll swap in other models later when a task needs more capability.

Troubleshooting

Message	Likely cause	Fix
Key loaded: False	Secret name doesn’t match	Name it exactly GEMINI_API_KEY (case-sensitive)
SecretNotFoundError	Notebook access off	Toggle Notebook access on for the secret
PermissionDenied / InvalidApiKey	Key invalid or expired	Generate a new key in AI Studio
ResourceExhausted / quota exceeded	Free-tier rate limit hit	Wait 1-2 minutes, then retry
ModelNotFound	Wrong model name	Use e.g. gemini-2.5-flash-lite
ModuleNotFoundError: google.genai	SDK not installed	Re-run !pip -q install -U google-genai

Revision summary

Must-know	One-line recall
Environment	Labs run in Google Colab (cloud notebooks) and call the Gemini API
API key = password	Identifies my account; whoever holds it can spend my quota
Safe storage	Put the key in Colab Secrets as GEMINI_API_KEY; never hard-code it
Load the key	userdata.get("GEMINI_API_KEY") → env var; genai.Client() reads it automatically
Install & call	!pip install -U google-genai; client.models.generate_content(model=..., contents=...)
If a key leaks	Delete it in AI Studio and make a new one - treat it as compromised