Providers
To connect a language model provider, run the command Posit Assistant: Configure AI Providers or select Configure Language Model Providers from the Accounts menu, then select Connect next to the provider.
For what each provider needs and how to sign in, see Providers in the Posit Assistant documentation.
| Provider | Chat | Code Completions |
|---|---|---|
| Posit AI Pass | ||
| Amazon Bedrock | ||
| Anthropic | ||
| Databricks Preview | ||
| DeepSeek | ||
| GitHub Copilot Preview | ||
| Google Gemini | ||
| Google Vertex AI (Gemini Enterprise Agent Platform) Preview | ||
| LiteLLM Preview | ||
| LM Studio (local) | ||
| Microsoft Foundry | ||
| Ollama (local) | ||
| OpenAI | ||
| OpenAI Compatible (legacy) | ||
| OpenCode Preview | ||
| OpenRouter (use a custom provider) | ||
| Portkey Preview | ||
| Snowflake Cortex |
If your provider is not listed, add a custom provider. It can use any of the provider types above, so you can connect a company gateway, a second account on the same provider, or a local model server.
providers.json
Posit Assistant saves your provider settings, but not your keys or tokens, to ~/.posit/ai/providers.json. Some changes are not on the AI Providers screen, such as turning a provider off, sending custom HTTP headers, or filtering the model list. For those, run Open AI Provider Settings (JSON). For every key the file accepts, see the providers.json reference.
Positron 2026.10 removed the authentication.*, positron.assistant.provider.*, and positron.assistant.models.overrides.* settings. If they are still in your settings.json, Positron moves them to providers.json on startup. To run this again, use Migrate AI Provider Settings to providers.json.