For developers
Once reporting is enabled, Sessions presents four tabs covering an individual's own work: My work, Activity, My report, and Costs. By default every tab is scoped to the signed-in user; a Me / Team toggle switches to the team's combined figures for anyone with visibility into others' data.
My work
A summary of lifetime activity: total tokens and sessions over the last 12 months, number of active days, current and longest streak, busiest day, and the most frequently used model. A year-long activity grid and a breakdown by tool follow, along with the current week's sessions, spend, and a day-by-day token chart, plus a table of the branches worked on recently and how much each one used.
Activity
A recent, moment-to-moment view of usage. Selecting a time window (24 hours, 7 days, 30 days, or all time) shows sessions per day, total calls and tokens, and the number of machines reporting, along with a table listing every individual session: start time, branch, machine, tool, model or models used, number of calls, tokens, and duration. Figures update live while the tab is open.
My report
A written summary of working style, based on five scored dimensions: steering, delivery, planning, quality, and design sense. These dimensions describe differences in working style rather than a scale from worse to better, and a radar chart plots them against what is typical for other people here. Below the summary are several specific figures, such as longest session, longest streak, and total time spent working with an agent, each with a brief explanation of what it measures.
Overview presents the summary itself, together with the score compared to what is typical here. Analysis shows the data behind it for each dimension: the score, what is typical, the contribution from subagents, and the actual terms from the person's own messages that moved the score; never a score with nothing behind it. It also breaks down when the person tends to work and how often they redirect an agent with a new direction rather than a bare stop, and keeps a history of every time the score was computed, so a figure can be traced back to the run that produced it.
Costs
Individual AI spend for the selected time window (7, 30, or 90 days, or all time): the total amount, the model accounting for most of it, a day-by-day chart broken down by model, and a breakdown of where the spend went. Below that are running totals for tokens in and out, how much of the input was served from cache, the number of AI responses, and a projected total for the current month based on the trailing week's rate.