Google Gemini

Connect Spaceduck to Google AI for Gemini chat models and text embeddings.

What it’s good for

Fast, capable models with generous free tiers. Gemini handles both chat and embeddings from a single provider, so there’s no two-server setup needed. Good for getting started quickly with a cloud provider.

Requirements

• A Google AI API key
• Generate one at aistudio.google.com

Configure in Spaceduck

Chat

1. Set your API key

   In Settings > Chat, select Google Gemini as the provider, then enter your API key.

   Or via CLI:

   spaceduck config secret set /ai/secrets/geminiApiKey
   
2. Select a model

   In Settings > Chat:

   • Provider: Google Gemini
   • Model: e.g., gemini-2.5-flash or gemini-2.0-pro

   Or via CLI:

   spaceduck config set /ai/provider gemini
   spaceduck config set /ai/model "gemini-2.5-flash"
   
3. Verify

   Send a message in the chat. You should see a streaming response.

Embeddings

Gemini provides text embeddings through the same API key.

1. Configure embedding model

   In Settings > Memory:

   • Toggle Semantic recall on
   • Provider: Google Gemini (or “Same as chat provider”)
   • Model: text-embedding-004
   • Dimensions: leave empty for the provider default

   Or via CLI:

   spaceduck config set /embedding/enabled true
   spaceduck config set /embedding/provider gemini
   spaceduck config set /embedding/model "text-embedding-004"
   
2. Test

   Click the Test button in Settings > Memory.

Test and troubleshoot

Problem	Cause	Fix
401 or 403 error	Invalid API key	Check your key at aistudio.google.com
429 Too Many Requests	Rate limit (free tier)	Wait a minute and retry, or upgrade to a paid plan
Model not found	Typo in model name	Check available models at ai.google.dev/models