Retain for data scientists
When a meeting ends with "can someone pull the numbers", Retain can answer with a notebook: Python, your data and charts, running on your own Mac. This page walks through a week of that, and says plainly what notebooks are and are not for.
Record the meetings that need numbers
Record the planning, product and review meetings where questions about data come up. Retain transcribes them with speakers. Meeting types lists the options.
Answer with a notebook
A notebook is made of cells you can run: Python, SQL, text and charts. You ask the chat to write and run cells, or write them yourself. Charts are drawn from query results, and a caption says where the data came from. Each cell can be edited, run, collapsed and moved. Notebooks explains the parts.
The code runs in a small virtual machine on your Mac, not in the cloud. This needs the Mac app on a Mac with Apple silicon running macOS 26 or later. On other Macs you can read and edit a notebook with its saved outputs but not run it. Notebook sandbox explains the isolation.
Bring your data in
You make one setup, under Settings and then Setup, and every notebook starts from it. You declare folders, which appear read-only inside each notebook, add the Python packages you want, and allow the hosts or database tunnels your code may reach. You can also attach a file in the notebook chat or drop a file into the notebook's work folder. The AI sees file names and sizes and reads values by running code, and Retain keeps no credentials. See notebook setup and data and charts and export.
Read next
Need help?
Can't find what you're looking for? Contact our support team for a quick reply.
Contact Support