Skip to content

Use isoboard with a Google Gemini API key

You can use isoboard with your own Google Gemini API key. Create a key in Google AI Studio, paste it into the Gemini row on the Providers tab, and pick a Gemini model in a Profile. isoboard stores the key in the macOS Keychain, and Google bills you directly where your plan charges for use.

  1. Open Google AI Studio’s API key page and create a key. Copy it.
  2. In isoboard, open Settings and select the Providers tab.
  3. Expand Gemini and paste the key into API key.
  4. Click outside the field. A status below it shows Ready or Configured. Configured means the key is saved and isoboard has not used it yet.

isoboard then creates two starting Profiles, Gemini - Efficient and Gemini - Accurate. Efficient favors faster, lower-cost captures. Accurate favors difficult captures and may take longer and cost more. You can edit or delete both. See Profiles.

Any Gemini model that accepts images. In the Profiles tab, choose Gemini as the Provider, then search the model picker by name or type a model ID. A model without image input fails at capture time. isoboard does not block it when you save.

In the macOS Keychain, never in the configuration file. isoboard does not show the key again after you save it, and you replace it by pasting a new one. See where isoboard stores your data.

You do. Each capture sends an image to Google, and Google applies its own pricing, free-tier limits and terms to your key. isoboard adds no fee. The isoboard license is separate. See pricing and licensing.

As of November 2026, the Gemini API has a free tier for some models. On the free tier, Google may use your prompts and responses to improve its products, and human reviewers may read them. Paid usage is not used that way. Do not send sensitive or confidential captures on a free-tier key. Check the current terms on Google’s Gemini API pricing page before you rely on this.

Message in isoboard Likely cause
the API key is invalid, or provider authentication is required The key was deleted, mistyped or restricted. Paste a new key.
provider access was denied Your project or region cannot use that model.
provider inference is rate limited You hit a quota. Wait, or check your plan.
inference timed out The request took longer than the Profile’s timeout.

More fixes: provider errors.