Model Library
The Model Library is where you manage your AI provider API keys and control which models appear in workspace dropdowns.
Go to Settings → Model Library.
Adding a provider
- Click Add provider
- Select your provider (Anthropic, OpenAI, Google, Groq, Together)
- Paste your API key
- Click Save
The platform validates the key and fetches the available models from that provider. They appear in the model list immediately.
Choosing which models appear
Each model in the library has a visibility toggle:
- Visible — the model appears in workspace creation and the model switcher
- Hidden — the model is removed from all dropdowns (the key is kept; you can unhide at any time)
Hide models you don't use to keep dropdowns clean.
Security
API keys are encrypted at rest — they are never stored as plain text and never appear in logs. The platform uses the key only at the moment of making an AI request.
If you suspect a key has been compromised, revoke it in your provider's dashboard and re-enter a new one in the Model Library.
Supported providers
| Provider | Models |
|---|---|
| Anthropic | Claude Haiku, Sonnet, Opus |
| OpenAI | GPT-4o, o1, o3 and others |
| Gemini Flash, Gemini Pro | |
| Groq | Llama 3.3, Mixtral and others |
| Together AI | Open-source model catalogue |
Operator models
The platform uses a separate pair of AI models to interpret the coding agent's terminal output and turn it into readable chat messages. These are configured in Settings → Operator Models and are separate from the model used for coding tasks.
See Operator Models for details.