For Statisticians ·
What you'll accomplish
Most code assistants only see the file you have open. Posit Assistant, built into Positron (Posit's IDE for R and Python), also picks up context from your loaded data, plots, and console history, so a suggestion can account for what you actually have in memory during an interactive session, not just the visible code. If you already work the way most statisticians do (load data, explore, plot, adjust) this guide sets it up around that exact rhythm.
What you'll need
What you should see: A familiar IDE layout: an editor pane, a console, and side panels for things like your environment and files.
Positron lists every supported provider by default, but none are active until you authenticate with one.
What you should see: A list of available providers, including options like Posit AI Pass, Anthropic, GitHub Copilot, Amazon Bedrock, and others, each showing how it authenticates (OAuth, an API key, or a managed credential depending on the provider). Troubleshooting: If a provider you expect is missing from the list, check whether your organization manages Positron centrally. Some providers are only available through Workbench-managed credentials.
What you should see: The provider now appears as active, no longer just listed as available.
What you should see: A response that references your actual loaded data or console history, not just the code visible in your editor, confirming the context-aware behavior is working.
Posit Assistant supports a few features worth knowing from the start.
What you should see: Each feature behaving as described: attached files feeding into the conversation, thinking effort changing how long a response takes, and Plan Mode producing a written plan before touching your code.