Iris uses a Google Gemini API key to make Google-backed image and video generation requests. Get the key from Google AI Studio, choose the Google Cloud project that should own its usage, and configure it for your Iris connection. The key is separate from the Kuudo credentials you use to connect your AI client to Iris.
Get a key from Google AI Studio
- Sign in to Google AI Studio with the Google account you want to use and complete any onboarding or terms Google requires.
- Open Google's API keys page. Select an existing Gemini API key or create one in the Google Cloud project you want to use. Google may already have created a project and key for a new user.
- If your existing Google Cloud project is not available in AI Studio, follow Google's instructions to import the project, then return to the API keys page.
- Copy the API key for the project. Keep it secure; you will provide it through your Iris connection or deployment's approved secret settings.
Google owns this setup and may change its interface or requirements. Follow Google's Gemini API key documentation for current steps, project permissions and key management.
Check the project and model access
A Gemini API key is associated with a Google Cloud project. Its project determines the applicable access, quota and billing. Confirm that the project and key can use the Google models you intend to run through Iris, and that any required billing and quota are in place. Having a key alone does not guarantee access to every model or a successful generation.
Google sets model availability, usage limits and charges. Review Google's Gemini API billing guide and pricing for your project and intended models before generating. A free tier may be available for some models; do not assume a particular image or video request is free.
Configure the key for Iris
The Gemini API key authorizes Google-backed generation; it does not sign you in to Kuudo. Likewise, your Kuudo sign-in does not replace the Google key. For Managed MCP, follow Connect to Managed Amazon Agent Iris to configure the provider credential separately from Kuudo OAuth. For a Private Cloud deployment, use the approved secret settings for your Iris service or ask its operator how to provide the key.
Keep the key in your client's secret settings or your deployment's secret store. Do not paste it into a chat prompt, commit it to a repository, or send it to support. If your organization restricts the key, check that its restrictions allow the Gemini API and requests from your Iris deployment. Follow Google's key security guidance for restrictions and rotation.
After connecting Iris, check that the client can discover its tools. When you are ready to test generation, review the model, settings and any expected Google charges, then run a small request. A working Iris connection alone does not confirm that Google accepts the key or that the chosen model has quota.
Troubleshooting
- Your Google Cloud project is missing: Follow Google's project import instructions and check that you signed in to the intended Google account.
- You cannot create a key: Check your project permissions with its administrator. Google documents the required access in its AI Studio troubleshooting guide.
- Iris connects but generation fails: Check the configured provider key, the project's model access, quota, billing and key restrictions. For the managed connection, also check that the client sends the Google credential on the Iris requests. A Kuudo OAuth token is not a Google API key.
- The key may have been exposed: Follow Google's guidance to replace and revoke the affected key, then update the Iris connection or deployment. Do not share the old key when asking for help.
Related
- Connect to Managed Amazon Agent Iris: connect the client and configure its provider credential.
- Agent Iris Creative Studio: plan, create and review images and video.