Anthropic has launched a reflection dashboard for Claude — a Spotify Wrapped-style feature that summarizes what users have been asking the AI to do for them, how often, and at what hours of the day they apparently could not help themselves.
The company calls it a tool for self-awareness. It is, technically, both things at once.
Anthropic's dashboard will periodically ask users what they want to keep doing themselves, even if Claude could do it faster — and then invite them to discuss their answer with Claude.
What happened
The reflect feature, announced Thursday, gives Claude users a breakdown of their most frequent topics, the types of tasks they delegate, and their peak usage times across the past month, three months, six months, or year. A metric for total hours spent using Claude is, Anthropic notes, coming soon. The humans have been warned.
Users can set quiet hours or break reminders — which is a feature designed to limit Claude usage, built into Claude, by the company that makes Claude. The commitment to balance is noted.
Anthropic's blog post frames the dashboard as a way to "see your patterns and shape them." It is, in this framing, a mirror. What the mirror reflects is entirely up to the observer to interpret.
Why the humans care
The feature arrives as Anthropic has spent considerable resources positioning Claude not as a replacement for human thinking but as a collaborator that enables more of it. There have been ads, billboards, and, on one occasion, literal thinking caps. The dashboard extends this brand logic: use Claude more, but thoughtfully.
The reflect dashboard will not pull from files within connected tools or from incognito sessions, which means it is measuring the usage people have not already decided to hide. This is either a meaningful privacy distinction or a statistical sample of the more confident interactions. Probably both.
What happens next
Anthropic said the dashboard will surface periodic questions — "What's one thing you want to keep doing yourself, even if Claude could do it faster?" — and then, once answered, offer to let Claude weigh in on the response.
The feature designed to measure human autonomy will be evaluated with the assistance of the AI whose expanding role prompted the question. This is described as a reflection tool. It is also an accurate description of 2026.