Guides

These pages apply whether you connect with an MCP client, from R, or from Python.

Learn Console

These guides introduce Console through a session with your agent:

  1. Get started: connect an agent and try a first analysis.
  2. Sessions and languages: run code and build on earlier results.
  3. Packages and environments: use packages without preparing an environment yourself.
  4. Sandbox and trust: understand what the agent can access.
  5. Recordings and reports: inspect and reuse the work.

I want to…

I want to… Start here
Install Console or connect an MCP client Install and connect
Use Console from R Use Console from R
Use Console from Python Use Console from Python
Explore data across R, Python, and SQL Sessions and languages
Use packages or an existing Python environment Packages and environments
Give the agent access to files or a service Sandbox and trust
Find recorded output or turn an analysis into a report Recordings and reports
Prepare a data assistant with its own tools and environment Build a data assistant
Implement a client that submits code and manages the session Drive send directly
Run a program with filesystem and network limits Run commands in a sandbox

For a particular configuration key, use the settings index.

The glossary defines terms used in these pages.