Usage Analytics

The Cursor Web Dashboard provides usage analytics so you can understand how your team is using Cursor.

Data Access and Visibility

Team admins have access to data for themselves and all other users in the team. Team members without admin privileges can see data for themselves and in some cases (like the Usage Leaderboard) for select other users on the team.

CSV Download

Each chart has a button on the bottom-right corner which allows for CSV download of visible data. Additionally, users can download data for all charts by clicking the download icon in the page header.

API Access

See our Admin API documentation to access analytics data programmatically. Available only for Enterprise customers.

Filtering Data

Dashboard users can filter usage shown for specific users, active directory groups, and dates via the header. Filtering supports up to 10 users and 90 continuous days of data.

Clicking on the gear icon in the header allows users to select timezone as well as whether weekends are shown.

Tracking AI Code in Git Commits

Cursor keeps a log of the signature of every AI line (Tab or Agent) that is suggested to the user during their chat session.

These lines are stored and later compared to the signatures of each line in subsequent git commits that were written by the same author. Cursor will detect all the line changes (additions or deletions) written by the Cursor Agent or Tab, and attribute the line as being written by AI.

All the AI detection is done on device, and never leaves the user's computer. We store the line counts as metadata and make them available via API or in the Analytics Dashboard.

Known Limitations:

• Diff signatures may be invalidated if automated code formatting is modifying lines.

• AI Code Tracking has not been implemented for Background Agents, or the Cursor CLI yet.

• All code signatures are stored on-device. The git commit must be scored on the same machine as the AI code was authored.

AI Output

AI Share of Committed Code

AI Share of Committed Code shows the lines of code changed in commits to your repositories, and what % of that code was generated by Cursor. Users can filter for production branch, which will use:

• The optional default branch set to the git repo

• Fallback to common default branch names such as: main, master, production, prod.

We use the following definitions:

• Cursor AI: Any line that can be attributed to Cursor Agent or Tab based on diff signatures.

• Other: Any line of code that can't be detected as being written by Cursor

Agent Edits

Agent Edits shows the amount of code edited by the Agent, and Cmd+K, and whether those changes were accepted by the user. Viewers can group the data by suggested / accepted or by file extension.

Tab Completions

Tab Completions shows the number of times Tab code has been suggested (and accepted) by users. The unit is Tab suggestions, regardless of lines of code changed in that suggestion.

Messages Sent

Messages Sent shows the number of messages sent by users to Cursor. Users can filter this data by the mode (e.g., Agent, Ask, Cmd+K) or by models used.

Active Users

The Active Users chart shows the number of unique active Cursor users in your team across different products. A user is defined as active in a period if they use at least one AI feature (Tab, Agent, Background Agent, CLI). Bugbot users are synced to Github accounts (not Cursor) and therefore not included in the All active user rollup.

Daily Usage

The Daily Usage chart visualizes Cursor activity over the preceding 365 days.

Users can toggle to see this view by:

• All: shows lines of code edits suggested by AI in Cursor (Tab and Agent).

• Tab: shows the number of suggestions made by Tab.

• Agent: shows lines of code suggested by Agent.

• DAU: shows daily active users across all Cursor products.

Usage Leaderboard

The Usage Leaderboard shows top Cursor users across your team alongside their favorite model and select usage stats for the selected time period.

We provide the following metrics:

• Chats: number of messages sent by the user in the chat interface (Agent, Plan Mode, Ask Mode, etc).

• Tab Completions: number of Tab suggestions accepted by the user.

• Agent Lines of Code: Lines of code written by the Agent and accepted by the user.

The top ten users and any filtered users are always shown. All users in the team are able to view the leaderboard.

Repository Insights

Repository Insights allow you to see how Cursor is used across different repositories. We report on:

• AI Lines of Code Committed: Code written by Cursor (Tab and Agent) that was committed by a user.

• Total Lines of Code Committed: All code committed by users.

• Code Committed by AI %: The % of lines of code committed that were edited by Cursor (Tab and Agent).

Some commits will be associated with Unknown repository if the user makes commits to a local git repository that doesn't contain a remote origin, or a remote that couldn't be resolved by Cursor

Conversation Insights

Cursor analyzes the code and context in each agent session to understand what kind of work is occurring. This makes Cursor the first self-aware software engineering platform, synthesizing the type of work happening across your team.

Teams no longer need high-toil, lossy analysis of tickets or low-response surveys to understand engineering work. Conversation Insights lets you deeply understand the type of work being done with Cursor. For work your team hands to Bots, see Grok Bot Conversation Insights.

Classification Dimensions

Conversation Insights classifies work across these dimensions:

• Category: Bug Fixing & Debugging, Code Refactoring, Code Explanation, Configuration, New Features, UI/Styling, Architecture, Data/Database, Documentation, DevOps/Deployment, Learning, Testing

• Work Type: Maintenance (KTLO), Bug Fixing, New Features

• Complexity: Distinguishes between the complexity of tasks teams assign to agents

• Specificity: Measures how specific the prompts developers use with agents are

Enterprise customers can extend these default categories or define their own across the organization or within specific teams.

Compare

Compare allows you to select and compare usage across teams and individual developers within your organization. Use this to identify adoption patterns, find power users, and understand how different groups use Cursor.

Privacy and Data Handling

All classification runs on-device. Default classifiers ensure no PII or sensitive data leaves the machine. The model outputs are validated against expected values. Any responses that don't match are discarded.

Pricing

Conversation Insights is free during the preview period. Starting January 1st, 2026, customers will be charged for inference. The Cursor Token Rate applies when the underlying request is a third-party model request subject to the rate, including when Auto routes to a third-party model.

Grok Bot Conversation Insights

Grok Bot Conversation Insights applies the same idea as Conversation Insights to the Bots your team runs. Cursor classifies each Grok Bot conversation so admins can see what kind of work Bots handle across the team, without reading the conversations themselves.

Bots handle office work rather than editor sessions, so the Grok Bot view uses its own dimensions:

Classification Dimensions

Grok Bot Conversation Insights classifies each conversation across these dimensions:

• Type of Work: The job the Bot handled. Inbox & Briefings, Calendar & Meetings, Sales & Pipeline, Customer Support, Marketing & Content, Finance & Back Office, People & Recruiting, Engineering & Product, Research & Documents, Errands & Purchases, Setup & Exploration, Other

• Level of Automation: How far the Bot acted on its own.

  • Not automated: The Bot answered or recommended. A person did the work.
  • Semi-automated with human review: The Bot drafted or staged a change and waited for approval.
  • Fully automated: The Bot completed the action in the target tool.

Compare

Compare works with the Grok Bot dimensions. Select teams or members to compare which jobs their Bots handle and how much of that work runs fully automated.

Privacy and Data Handling

Grok Bot Conversation Insights reports aggregate counts per dimension. The dashboard doesn't show or export conversation content. Your team's Privacy Mode setting governs whether Grok Bot conversation data is used for training. See Grok Bot security.

Cloud Agent Usage

Agents Created

Agents Created shows Cloud Agent usage by the originating source. Each time a Cloud Agent starts up counts as one Agent.

Pull Requests

Pull Requests shows Pull Requests Opened and Merged that originate from Cloud Agents.

Lines of Code

Lines of Code shows code written and merged by Cloud Agents.

Cloud Agent Top Repositories

Cloud Agent Top Repositories shows repositories by number of Pull Requests opened and merged.

Top Cloud Agent Users

Top Cloud Agent Users shows top users by number of Agents Created. Viewers can also view data by Pull Requests opened and merged.

Client Versions

Client Versions shows which versions of the Cursor editor your team is using. Each user's version is the Cursor version they last opened during that day.