Connect a provider
BIGGI needs a model provider before it can run an agent task. The current release connects directly to a provider you choose; a BIGGI account is not required.
Connect in VS Code
- Open the BIGGI sidebar.
- Select the gear icon.
- Open Providers.
- Choose a provider.
- Enter an API key or complete the provider login flow.
- Open Models and select a model from the connected provider.
The connection and model list are shared by the sidebar, editor tabs, and Agent Manager.
Common choices
| Provider | Use it when |
|---|---|
| Anthropic | You want Claude models with an Anthropic API key. |
| OpenAI | You want GPT or Codex models with an OpenAI API key. |
| ChatGPT Plus or Pro | You want to use a supported ChatGPT subscription login. |
| Google Gemini | You want Gemini models through Google AI. |
| OpenRouter | You want one key for models from multiple labs. |
| Ollama | You want to run compatible models locally. |
BIGGI also discovers providers from supported environment variables and local configuration. See AI Providers for more connections.
Provider keys and privacy
Credentials are stored by the local BIGGI runtime. Do not commit provider keys to the repository. Model prompts and context leave your machine when you use a hosted provider, so review that provider's data policy before sending sensitive code.
Networked tools have their own boundaries. Web search, browser automation, remote MCP servers, and session sharing can also send data outside the workspace.
Autocomplete uses its own provider setting
Chat and autocomplete can use different models. Open Settings → Autocomplete and choose a supported autocomplete provider and model. If the selected autocomplete provider is not connected, inline suggestions remain idle.
Continue with Your First Task.