Cursor analytics
for engineering teams.

How to check your Cursor usage

On The X Subscription, your usage is under Usage at grok.com/dashboard, and in the editor's settings: how much of your weekly pool for Cursor's own models (Grok and Composer) you've used, the next reset date and your prepaid credit. Third-party models such as Claude and GPT aren't in the pool; they bill from that credit.

On a legacy plan (Hobby, Pro, Pro+, Ultra, or a Teams Standard or Premium seat), the Spending tab at cursor.com/dashboard/spending shows two monthly pools, Cursor Models and Other Models, and any on-demand charges. Either way, Cursor publishes no number for a plan's included usage; what each plan includes is on Cursor pricing. To price a month of tokens yourself, use the token cost calculator.

Who needs more than Cursor's own dashboard?

If your team only runs Cursor, its dashboard is enough. It counts Tab, agent edits and messages per person, and Zest doesn't record Tab at all. You need more when the same engineers also run Claude Code or Codex, or when you want to see which local sessions ended in a merged pull request, because Cursor's analytics docs list pull requests for cloud agents only.

We built Zest for the lead whose team mixes Cursor with other coding agents and keeps its code on GitHub. One developer checking a quota already has the answer in the steps above, and a team that needs Tab counts, or hosts its code elsewhere, should look somewhere else.

How to track Cursor usage across your engineering team

Start with Cursor's analytics dashboard, on a Teams or Enterprise plan with everyone on Cursor 1.5 or later: admins see each user's usage there. Enterprise teams can pull the same numbers from the Admin and Analytics APIs.

All of these cover Cursor only. To see Cursor and Claude Code usage in one dashboard, you either load both vendors' exports into your own reports, or use a tool that reads both agents, such as Zest.

What does the Cursor analytics dashboard show?

It shows how much of your committed code Cursor generated, along with agent edits, Tab completions, messages, active users and a leaderboard. On Teams and Enterprise, cursor.com/dashboard/analytics gives admins every user's data and members their own. Data comes only from Cursor 1.5 or later, and you can filter by up to 10 users and 90 days or download any chart as a CSV.

• AI share of committed code: the lines changed in commits, and the share Cursor generated.
• Agent edits and Tab completions, suggested and accepted.
• Messages sent, by mode (Agent, Ask, Cmd+K) and by model.
• Active users, and daily usage over the past 365 days.
• A usage leaderboard lists the top ten users with their favorite model, chats, Tab completions and agent lines of code, and every member can see it.
• Repository Insights gives AI lines committed, total lines committed and the share committed by AI, per repository.
• Cloud agent usage counts agents created and the pull requests they opened and merged, and ranks the top repositories and users.
• Conversation Insights, on Enterprise, classifies agent sessions by category, work type, complexity and specificity.

How do I pull usage and seat data from the Cursor Admin API?

An admin creates a key with the admin:* scope on the API Keys page of the Cursor dashboard, and you call https://api.cursor.com with it as the basic-auth username. Cursor's API overview lists the Admin API for Enterprise teams. Four endpoints cover usage and seats:

• POST /teams/daily-usage-data: per user and day, lines added and accepted, Tab accepts, agent, chat and Cmd+K requests, and the most used model.
• POST /teams/filtered-usage-events: each request with its model, its input, output and cache tokens, and its cost.
• POST /teams/spend: each member's spend in the current billing cycle. spendCents counts on-demand use only, and overallSpendCents adds included usage.
• GET /teams/members: each member's name, email, role and whether they have been removed, for a seat count.

curl -X POST https://api.cursor.com/teams/daily-usage-data \
  -u YOUR_API_KEY: \
  -H "Content-Type: application/json" \
  -d '{
    "startDate": 1710720000000,
    "endDate": 1710892800000,
    "page": 1,
    "pageSize": 1000
  }'

Is there a Cursor analytics API?

Yes, for Enterprise teams. The Analytics API returns the dashboard's numbers as JSON, with team totals under /analytics/team/ and one row per person under /analytics/by-user/. It covers agent edits, Tab use, daily active users, models, MCP servers, Cursor skills, commands and the leaderboard, up to 30 days per call. The AI Code Tracking API, in alpha, goes down to each commit: lines from Tab, from the agent and from people.

Cursor counts AI lines per commit, GitHub Copilot counts them in the editor and Claude Code's contribution metrics in merged pull requests, so their shares don't add up. How much code is written by AI sets them side by side.

curl -X GET "https://api.cursor.com/analytics/team/dau?startDate=14d&endDate=today" \
  -u YOUR_API_KEY:

curl -X GET "https://api.cursor.com/analytics/by-user/agent-edits?page=1&pageSize=50" \
  -u YOUR_API_KEY:

How do I see Cursor and Claude Code usage in one dashboard?

Install Zest's extension in Cursor and its plugin in Claude Code, and each engineer's sessions from both tools land on one dashboard, next to the pull requests they merged. Cursor's dashboard and APIs stop at Cursor, so they can't show the Claude Code or Codex sessions the same engineer runs. Each engineer adds the Cursor extension with one click (setup). It reads Cursor's local chat history (composer conversations and agent transcripts) and syncs each session with its models and timings; optionally, it adds each session's tokens and cost from Cursor's usage dashboard.

Cursor is the agent our own team runs least: it co-authored 7 of the 3,192 commits in our two repos from July 5 to October 3, 2026. What this page says about Cursor comes from Cursor's docs and the extension's code, not from our usage.

• MCP tools and the skills someone asks for by name in each Cursor session, rolled up per engineer in the AI Stack view.
• Cursor next to Claude Code, Codex and Copilot on the same dashboard, so Claude Code vs Cursor or Cursor vs GitHub Copilot is answered on your own team's work.
• With the Zest GitHub App, each engineer's Cursor sessions sit next to the pull requests they opened and merged.
• Redaction runs on the engineer's machine before upload; no code diffs are sent.