Skip to main content

Documentation Index

Fetch the complete documentation index at: https://mintlify.com/derailed-dash/gemini-file-search-demo/llms.txt

Use this file to discover all available pages before exploring further.

Every agent in this codelab reads its configuration from environment variables at startup. Setting these variables correctly — particularly GEMINI_API_KEY — is a prerequisite for running any of the SDK or ADK agents. This page walks you through obtaining an API key, creating your .env file, and loading the variables into your shell.

​
Environment variables

VariableRequiredDefaultDescription
GEMINI_API_KEYYes—Your Gemini API key from Google AI Studio. The SDK reads this automatically via genai.Client().
MODELNogemini-2.5-flashThe Gemini model ID to use across all agents.
STORE_NAMENodemo-file-storeThe display name of the File Search Store to attach to RAG agents. Must match the name you used when creating the store in the notebook.
GEMINI_API_KEY must be set before running any agent. Every agent creates a genai.Client() on startup, which requires the key to be present in the environment.

​
Getting a Gemini API key

The Gemini File Search tool requires the Gemini Developer API, which uses an API key rather than service-account credentials.
1

Open Google AI Studio

Go to Google AI Studio and sign in with your Google account.
2

Create an API key

Follow the steps in the API key guide to generate a key for your Google Cloud project.
3

Copy and keep the key safe

Copy the key immediately after it is displayed. Store it somewhere secure — treat it like a password.

​
Creating your .env file

The repository ships with .env.template as a starting point. Copy it to .env and fill in your values.
1

Copy the template

cp .env.template .env
2

Edit the file

Open .env in your editor. The template contains:
export GEMINI_API_KEY="your-api-key"
export MODEL="gemini-2.5-flash"
export STORE_NAME="demo-file-store"
Replace your-api-key with your actual Gemini API key. Adjust MODEL and STORE_NAME if needed.
3

Load the variables

Run the following command in the same terminal session you will use to run the agents:
source .env
You must re-run source .env in any new terminal session before running an agent.
.env is listed in .gitignore. Never commit a file containing a real API key to version control.

​
What happens when STORE_NAME is not set or the store is not found

Both sdk_rag_agent.py and rag_agent_adk/agent.py perform a graceful fallback rather than crashing when the File Search Store is unavailable.
  • If STORE_NAME is not set at all, the agents log: 
    WARNING - STORE_NAME env var not set. Skipping store lookup.
    and disable RAG capabilities for that session.
  • If STORE_NAME is set but no store with that display name exists, the agents log: 
    WARNING - Store '<name>' not found! RAG capabilities disabled.
    and continue running without the File Search tool attached.
If the RAG agent starts successfully but cannot answer questions about the story, check that STORE_NAME matches the display name you used when creating the store in notebooks/file_search_store.ipynb. Display names are case-sensitive.

​
MODEL variable

All agents fall back to gemini-2.5-flash when MODEL is not set. To use a different model, set the variable before sourcing: 
export MODEL="gemini-2.5-flash"
source .env
The model must support the tools you intend to use. File Search and Google Search are supported on current Gemini 2.x models. If you see a “model not found” error, verify the model ID is correct.

Build docs developers (and LLMs) love

Get started for freeTalk to us