Models & Providers
Choose a model provider that fits your privacy, latency, capability, and budget needs.
Provider options
CODVIAN can work with local and hosted model providers. Ollama is useful for local workflows, while OpenAI, Anthropic, Google, OpenRouter, and gateway integrations provide hosted options.
- Ollama for local models
- OpenAI-compatible APIs
- Anthropic Claude models
- Google Gemini integrations
- OpenRouter and managed gateways
Configure a provider
Open Settings, choose Models & Providers, and add the connection details requested by the provider. Keep keys in the place recommended by the provider and avoid sharing them in project files or session messages.
Capability support varies
Context length, tool use, multimodal input, streaming, and reasoning behavior depend on the selected model and endpoint. Test a small task before a large project.
Switch providers
You can change the active provider between sessions or tasks. Keep the task instructions provider-neutral when possible so the same workflow remains portable.