Claude Reflect: see how you actually use AI

Claude Reflect is a monthly recap of your chats: top topics, peak hour, an AI-skills breakdown. Here's how to turn it on and read it.

Claude Reflect (Anthropic calls it "Reflect," and its help article says "see your monthly recap") is a built-in dashboard that summarizes how you've been using Claude: your most active day, your peak hour, key topics, and a skills breakdown. It's in beta as of July 9, 2026, and it needs Memory turned on to generate anything at all.

What Reflect actually is

Reflect lives at Settings > Reflect on Claude web and Claude Desktop (or go straight to claude.ai/settings/reflect). Open it and Claude generates a recap on the spot, it isn't precomputed in the background, so the first load can take a moment.

You pick a time range from a dropdown: past month (the default), this month so far, past 3 months, past 6 months, or past year. Whatever window you choose, the recap only reflects it, per Anthropic's help center.

Why it exists. Anthropic frames Reflect as a mirror, not a scoreboard: a way to notice patterns in how you delegate to AI, not just how much you use it. Worth knowing going in: TechCrunch's launch-day coverage argues that visualizing how embedded Claude already is in your day doubles as a nudge to keep using it more. Both things can be true at once.

It's part of a small pattern in consumer AI right now. OpenAI shipped a similar "Your Year with ChatGPT" recap in December 2025, Spotify-Wrapped-style, for Free/Plus/Pro users with memory on. Reflect is Anthropic's version, minus the awards and the AI-generated poem, plus a wellbeing layer OpenAI's version didn't have.

How to turn it on and read your first recap

now

Turn on Memory

Go to Settings > Memory and toggle it on (or check Settings > Capabilities for "Generate memory from chat history"). Reflect only generates a recap when Memory is on.

now

Open Settings > Reflect

Or go directly to claude.ai/settings/reflect on web or Desktop. The recap builds live when you land on the page.

now

Pick your time range

Default is "past month." Switch to "this month so far," 3 months, 6 months, or past year from the dropdown if you want a wider or narrower slice.

Read each panel in order

Activity chart first, then key topics, then the 4D skills breakdown, then wellbeing tools. Each one answers a different question about your usage.

How to read each panel

Most active day and peak hour. A daily activity chart plus a called-out "most active day" and "peak hour," per the Help Center. This is the part that's closest to a literal usage log: when you actually open Claude, not what you use it for.

Total conversations. A raw count for the selected window. Useful mainly as a baseline before you look at the breakdown panels, since a high count with narrow topics reads differently than the same count spread across a dozen use cases.

Key topics. A summary of the subjects and task types your conversations clustered around. Anthropic's own copy says "key topics," not "top topics," so treat the exact ranking logic as unstated rather than a strict leaderboard.

The 4D AI-Fluency framework. Four dimensions Anthropic uses to describe how you engage with AI, worded exactly as:

Delegation: setting goals and deciding whether and how to engage with AI
Description: effectively describing goals to prompt useful AI behaviors and outputs
Discernment: accurately assessing the usefulness of AI outputs and behaviors
Diligence: taking responsibility for what we do with AI and how we do it

Read this panel as a self-assessment prompt, not a grade. It's meant to surface which of the four you lean on already and which you might be skipping, like whether you check Claude's output (discernment) as often as you ask it to produce something (description).

Wellbeing tools. Quiet-hours scheduling, a break nudge you can set after a chosen amount of use, and periodic reflective prompts, one example Anthropic gives is "What's one thing you want to keep doing yourself, even if Claude could do it faster?" Anthropic says these were built with input from MIT Media Lab's AHA program, Boston Children's Hospital's Digital Wellness Lab, and the Family Online Safety Institute.

“A view of how much time you've spent with Claude is on Anthropic's roadmap, but it's not live in this beta.”

Honest limits

Beta, rolling out gradually. Not every eligible account will see it on day one.
Free, Pro, and Max only. Team and Enterprise plans are explicitly excluded at launch.
Not on Claude Mobile. Reflect only shows up on Claude web and Claude Desktop.
Memory required. No Memory, no recap, full stop.
Claude Code and Claude Cowork activity isn't counted. Only regular Claude.ai and Desktop chat activity feeds the recap.
No time-saved metric yet. Reflect doesn't calculate hours saved. Anthropic has signaled that's coming later, not live now.
Privacy carve-outs. Incognito chats are excluded entirely. Connected-tool chats (like an inbox summary) can appear, but the underlying emails or files don't. Chats using a health integration such as Apple Health are skipped from insights entirely, and sensitive topics surface only as high-level summaries, never raw detail.

If Reflect isn't showing up, Anthropic's own FAQ starts with the same checks above: Memory toggled on under Settings > Capabilities, and a Free, Pro, or Max plan on web or Desktop.

What is Claude Reflect?

Reflect is a built-in Claude feature that generates a recap of your recent usage: activity patterns, key topics, and a 4D breakdown (Delegation, Description, Discernment, Diligence) of how you engage with AI, alongside optional wellbeing tools like quiet hours and break nudges.

How do I turn on Reflect in Claude?

Turn on Memory in Settings > Memory first, then go to Settings > Reflect (or claude.ai/settings/reflect) on web or Desktop. The recap generates live when the page loads; no separate opt-in toggle exists beyond Memory.

Does Claude Reflect require Memory to be turned on?

Yes. Reflect has no data to summarize without Memory on. If Memory is off, the recap won't generate at all, regardless of how much you've chatted with Claude.

Is Claude Reflect free?

Yes, it's included on Free, Pro, and Max plans at no extra cost, all in beta. It's not available on Team or Enterprise plans as of this writing.

If you're just getting oriented, how to use Claude covers your first session end to end. And once Reflect shows you where your usage actually goes, Claude Projects is the next step for turning repeat chats into a workspace Claude remembers on its own.