Three kinds of Claude Code dashboard, and the question each one answers

Search for a Claude Code dashboard and you will land on three different things. For an admin it means Anthropic's analytics pages for Team, Enterprise and API organizations. For a developer at the keyboard it means /usage and the status line inside the terminal. For someone on a Pro or Max plan who keeps running into limits, it usually means a desktop app such as UsageAtlas that holds one person's limits and history in view. As of November 2026, none of them replaces the others.

The dashboards Anthropic runs for organizations

Anthropic's analytics documentation describes two hosted dashboards, one per kind of account.

The Team and Enterprise page tracks lines of code accepted, suggestion accept rate, and daily active users and sessions. With the Claude GitHub app installed, it adds contribution metrics, still in public beta, that count merged pull requests containing Claude Code output, plus a top-ten leaderboard and a CSV export. Anthropic calls these numbers deliberately conservative.

The Console version swaps contribution data for money: daily spend and a per-member table, 30 days by default, with spend and accepted lines per person. Anthropic labels those spend figures estimates and points to the Cost page for actual charges.

Both serve managers asking whether the rollout pays off, not a developer asking how much of this week's allowance is left. For per-user tokens and cost across an organization, Anthropic points to OpenTelemetry export, which streams Claude Code metrics and events into a company's own monitoring backend.

The dashboard inside the terminal

For the person typing, Claude Code ships its own views. /usage, which /cost and /stats now open too, shows session cost, plan limit bars and activity stats. On Pro, Max, Ultra, Team and Enterprise plans it adds a breakdown of what is eating the allowance, by skill, subagent, plugin and MCP server, with suggestions, over the last 24 hours or 7 days. /insights writes an HTML report on recent sessions, including where they hit friction.

The status line keeps a few of those numbers on screen permanently. A script you choose receives the session's data, including five-hour and weekly limit percentages on Pro, Max and Ultra plans, and community tools like ccstatusline turn that into a configurable bar.

This layer has hard edges. /usage reads local session history on one machine, so work on a laptop and a desktop never adds up. Claude Code deletes local session files after 30 days by default. And all of it disappears when the terminal closes.

The personal dashboard in between

That leaves a gap for solo developers: a dashboard outside the terminal that keeps more than a month and covers more than Claude.

UsageAtlas is the free, open-source answer we would point most readers to. It is a local-first desktop app for Windows, macOS and Linux that shows Claude Code's five-hour and weekly windows with reset times, tokens and estimated cost by day and model, and history you can keep with an optional free account. It sends a desktop alert when a window drops below a threshold you set, and it puts Codex, Cursor, GitHub Copilot and other coding tools on the same screen. It does not read prompts or responses, and its cost figures are estimates rather than a bill. The project's own guide to pairing a status line with a usage dashboard covers how the terminal and desktop views divide the work.

Which one to open

• Leading a team on Team or Enterprise: claude.ai analytics for adoption, the spend report or OpenTelemetry for per-person cost.
• Paying through the API: the Console dashboard, then the Cost page for real charges.
• One developer on Pro, Max or Ultra: /usage for right now, UsageAtlas for the week, the month and the other tools.

If your company already collects Claude Code telemetry centrally, that pipeline is the better team view, and a desktop tracker adds little beyond personal alerts.