How-To Guide
Configure Google Gemini as Your Hermes Provider
Configure Hermes Agent to use Google's Gemini models directly via Google AI Studio — fast, capable, and cost-effective.
Quick answer
Configure the native Google Gemini provider for fast, cheap, high-context models — Gemini Flash is strong for high-volume tasks. Native integration means a direct API key, no middleman, and lower latency than routing through an aggregator. Set the Gemini key in config and pick a model.
Google Gemini offers excellent performance at competitive prices. Gemini Flash is particularly fast and cheap — great for high-volume tasks. Native integration means no middleman, lower latency, and direct access to Google's latest models.
$49/mo or $39/mo billed yearly · AI included · 14‑day guarantee
Before you start:
- ☑Hermes Agent installed
- ☑Google account
- ☑Google AI Studio API key
Steps
- 1
Get your API key
Go to aistudio.google.com/app/apikey and create an API key
- 2
Add the key to .env
Add GOOGLE_API_KEY=your_key_here to ~/.hermes/.env
- 3
Set the provider
In config.yaml, set model: provider: gemini
- 4
Choose a model
Set model: default: gemini-3-flash (fast) or gemini-3-pro (powerful)
- 5
Test the connection
Run hermes and send a test message
Pro Tips
- 💡Gemini Flash is one of the best auxiliary models — fast and cheap for vision, summarization, and compression
- 💡The GEMINI_API_KEY alias also works if you prefer that name
- 💡Gemini models have excellent context windows (1M+ tokens for some variants)
- 💡Use gemini-3-pro for complex reasoning, gemini-3-flash for speed
Troubleshooting
❌ API key rejected
✅ Verify the key at aistudio.google.com. Ensure you're using a valid API key, not an OAuth token.
❌ Model not found
✅ Check the exact model name. Use 'hermes model' to see available Gemini models.
❌ Rate limited
✅ Google AI Studio has free tier limits. Consider upgrading or using OpenRouter for higher quotas.
FAQ
How do I use Gemini with Hermes?
Set up the native Gemini provider with your API key and choose a model. Native means a direct connection — lower latency and no aggregator markup.
Which Gemini model should I pick?
Gemini Flash is fast and cheap, good for high-volume work; the larger Gemini models give more capability and very large context for complex sessions.
Native provider or Code Assist OAuth?
Use the native API-key provider for simplicity. Use the Code Assist OAuth path if that's the access you have; both reach Gemini, just via different auth.