
Access Controls and Collaboration
Work as a team. Understand API key scopes, sharing prompts, and managing permissions in a collaborative AI project.
Access Controls and Collaboration
AI development is rarely a solo sport. You need to share your prompts with designers and your code with engineers.
Sharing Prompts
Since AI Studio prompts are saved to Google Drive, you can Share them just like a Google Doc.
- Button: Click the "Share" icon in the top right.
- Link Sharing: You can generate a link. Anyone with the link can open the prompt in their AI Studio.
- Note: They cannot edit your original file unless you give them Edit access. Usually, they will "Make a Copy" of your prompt to experiment.
API Key Security
- Keys are scoped to the Project: An API key has full access to the Gemini API for that GCP project.
- Restrictions: In the Google Cloud Console, you can (and should) restrict API keys to:
- IP Addresses: Only your office server IP.
- Specific APIs: Only "Generative Language API", not "Google Maps" or "Compute Engine".
Team Workflow
- Prompt Engineer: Designs the prompt in Studio. Saves it. Shares the link with the dev.
- Developer: Opens the link. Clicks "Get Code." Copies the Python/JS snippet into the app repo.
- Deployment: The app runs using a service-account or secured API Key.
Summary
- Share prompts via Drive links.
- Restrict API keys in Cloud Console.
- Separate "Prototyping" (Studio) from "Production" (Codebase).
Module 3 Complete! You have mastered the Studio interface. In Module 4, we dive into the art of talking to the machine: Prompting Gemini Models.