Configure an AI Model

Before the agent can do anything, you must configure an AI model. This is the first thing to do after installing.

1. Open the sidebar

Press Ctrl+Shift+A (Cmd+Shift+A on macOS) to toggle the AI sidebar.

This is the global shortcut to show/hide the agent sidebar. You can see it any time under Settings → Shortcuts. If the sidebar does not open, click the NevoFlux sidebar button in the toolbar, or press the shortcut again to toggle it.

On first run — when no provider is configured yet — the sidebar shows a Welcome to NevoFlux screen with a Start Setup button that takes you straight to the model settings.

2. Choose a provider and add your key

The setup button opens nevoflux://settings/llm (the AI Models tab of Settings). You can also open it any time with Ctrl+Shift+A → the gear icon, or by visiting nevoflux://settings and choosing AI Models.

In the AI Models tab:

1. Click the card for the provider you want (e.g. Anthropic, OpenAI, DeepSeek, Qwen, …).
2. Paste your API Key.
3. (Optional) set a specific Model and a custom Base URL.
4. Tick Set as active provider.
5. Save.

Saving writes the provider's settings into your config.toml ([llm.<provider>]) and marks it active. The sidebar detects the new provider and switches from the welcome screen to the chat view automatically.

See Configuration → Providers & API keys for the full list of supported providers, their default models, and the environment variables you can use instead of storing keys in the file.

3. You're ready

Open the sidebar and ask the agent something. Continue to Start a conversation.

No keyring. API keys are read from your config.toml or from environment variables — NevoFlux does not store them in an OS keychain. Treat config.toml as a secret file.