Secure your Gemini credentials and prep the workspace

Set aside 5–10 minutes to gather credentials and clone the starter project. You will finish with a working local checkout and secrets stored securely.

Create a Gemini API key

1. Visit Google AI Studio. Sign in with the Google account that has Gemini access.
2. Open API Keys → Create API key. Choose Gemini API (not Vertex AI) to stay on the free quota.
3. Copy the generated key to a safe location. You can rotate it later if needed.
4. (Optional) If your organisation mandates Vertex AI, note the project/region—this codelab defaults to direct Gemini API calls.

Tip: Keys created in AI Studio are tied to your Google account. Rotate them periodically and never commit them to Git.

Clone the multi-agent project

Pick a workspace directory and clone the repository:

Copy
git clone https://github.com/gourav221b/trip-planner-agent-adk.git
cd trip-planner-agent-adk

The repository ships with two agents:

• personal_assistant: a travel orchestrator that coordinates weather, safety, and itinerary specialists.
• celebration_planner: a celebration ideation demo that sparks parallel brainstorming.

You will explore both throughout the codelab.

Configure environment variables

Create a dedicated environment file so ADK can read your Gemini credentials:

Copy
cd personal_assistant
cp .env.example .env

Open .env and set the keys:

Copy
GOOGLE_GENAI_USE_VERTEXAI=0
GOOGLE_API_KEY="paste-your-ai-studio-key"

Return to the repository root:

Copy
cd ..

Safety check: ensure .env stays ignored by Git. The repository .gitignore already excludes it.

(Optional) Verify shell variables

If you prefer exporting variables directly (for CI or system-wide use), source the .env file:

Copy
set -a
source personal_assistant/.env
set +a
echo "$GOOGLE_API_KEY"

You are ready to install dependencies and boot the Google Agent Development Kit.