cursor-usage

Demo animation created with remotion-readme-kit
What This Does
You manage a Cursor Enterprise team. You want to know who's spending what, which models are being used, and whether you're getting value from AI. Instead of logging into a dashboard, you ask your AI agent:
"How much did my team spend this week?"
"Who's using the most expensive models?"
"Give me a full usage report for the current billing cycle."
This MCP server plugin wraps the full Cursor Enterprise Admin and Analytics APIs, giving your agent the tools and knowledge to answer those questions directly. Works with Cursor, Claude Code, and any MCP-compatible client.
Part of the cursor-usage-tracker ecosystem. This plugin handles quick questions in the IDE. The tracker is the full open-source dashboard with charts, three-layer anomaly detection, Slack/email alerts, and incident lifecycle tracking. Use them together or separately.
Related MCP server: Token Analyzer MCP
What's Included
Component | What it does |
MCP Server | 15 tools wrapping the full Cursor Enterprise Admin + Analytics API |
Skills | Data interpretation guide + cost optimization framework |
Rules | Always-on cost-awareness guidance for model selection |
Commands | Quick-access: |
Agent | Specialized usage analyst persona |
Install
Cursor (Marketplace)
/add-plugin cursor-usageThen set your API key in Cursor settings:
Open Settings → MCP
Find
cursor-usageand setCURSOR_API_KEYto your Cursor Enterprise Admin API key
Claude Code
/plugin install cursor-usageManual (any MCP-compatible client)
Add to your MCP configuration:
{
"mcpServers": {
"cursor-usage": {
"command": "npx",
"args": ["-y", "cursor-usage-mcp"],
"env": {
"CURSOR_API_KEY": "your-api-key-here"
}
}
}
}As a Skill Only (no MCP server)
Copy the skills/ directory into your project's .cursor/skills/ or .claude/skills/ folder. The skills work standalone as reference material even without the MCP server.
Getting Your API Key
Go to your Cursor team settings (Settings → Team → API Keys)
Generate an Admin API key
For analytics endpoints, you may also need an Analytics API key
The API key gives read access to your team's usage data. The only write operation is set_spend_limit.
Available Tools
Admin API
Tool | Description |
| List all team members with roles and status |
| Current billing cycle spend per member |
| Daily usage data: lines, requests, models, modes |
| Billing groups with member lists and spend |
| Per-request events with model, tokens, and costs |
| Set a hard spending limit for a user |
Analytics API
Tool | Description |
| Daily active users (includes CLI, cloud agent, Bugbot) |
| Model usage breakdown per day |
| Agent edit acceptance/rejection rates |
| Tab autocomplete effectiveness |
| MCP tool adoption |
| Top file types being edited with AI |
| Cursor version distribution |
| Command usage analytics |
| Plan mode adoption |
Composite Tools
Tool | Description |
| One-call summary: members, spend, DAU, top models |
| Deep dive into a specific user's usage patterns |
Example Conversations
Quick spend check:
You: "How much has my team spent this billing cycle?" Agent: calls get_team_overview → "Your team of 47 active members has spent $3,842 this cycle. Top spender is Alice at $412, followed by Bob at $287..."
User investigation:
You: "Why is Bob's spend so high?" Agent: calls get_user_deep_dive → "Bob has made 847 requests this week, 73% using claude-opus-4.5. His daily average is $41 vs the team median of $12. Switching his chat requests to Sonnet would save approximately $180/month..."
Model audit:
You: "/model-audit" Agent: calls get_model_usage + get_spending + get_agent_edits → "62% of your team's messages use Sonnet (good). However, 5 users account for 78% of all Opus usage. Their acceptance rate on Opus is 44% vs 51% on Sonnet, suggesting Opus isn't providing measurably better results for most of their tasks..."
Full Cursor Enterprise Cost Monitoring Dashboard
This plugin answers quick questions in the IDE. For teams that need:
Historical trends over months (API is limited to 30 days)
Automated anomaly detection with statistical outlier detection
Slack/email alerts when spend spikes
Incident tracking with MTTD/MTTI/MTTR lifecycle
Web dashboard with charts and visualizations
See cursor-usage-tracker, the open-source Cursor Enterprise dashboard this plugin is part of.
Development
# Install dependencies
npm install
# Run in development
npm run dev
# Build
npm run build
# Type check
npm run typecheck
# Run tests
npm test
# Validate plugin structure
npm run validateRequirements
Node.js 20+
A Cursor Enterprise team with API access
Admin API key (and optionally Analytics API key)
Author
Ofer Shapira
License
MIT © Ofer Shapira
Available Tools
17 toolsget_agent_editsB
Get agent edit metrics: suggested vs accepted vs rejected diffs and lines. Shows how effectively the team is using AI-generated code.
Inputs: startDate, endDate, users.
get_billing_groupsA
Get billing groups with member lists, group-level spend, and daily spend breakdown. Also returns billing cycle dates.
Inputs: no named parameters.
get_client_versionsA
Get Cursor client version distribution across the team: which versions are in use and what percentage of users are on each.
Inputs: startDate, endDate, users.
get_commandsC
Get command usage analytics: which Cursor commands are being used and how often.
Inputs: startDate, endDate, users.
get_daily_usageB
Get daily usage data per user: lines added/deleted, applies, accepts, rejects, tabs, requests by mode (composer/chat/agent), models used, and client versions.
Inputs: startDate, endDate, page, pageSize, +1 more.
get_dauB
Get daily active users over time, including breakdowns for CLI, cloud agent, and Bugbot usage.
Inputs: startDate, endDate, users.
get_file_extensionsB
Get top file extensions being edited with AI: which file types get the most AI suggestions, accepts, and rejects.
Inputs: startDate, endDate, users.
get_mcp_usageB
Get MCP (Model Context Protocol) tool usage: which MCP servers and tools are being used, and how often.
Inputs: startDate, endDate, users.
get_model_usageA
Get model usage breakdown per day: which models are being used, how many messages, and by how many users. Essential for understanding model adoption and cost drivers.
Inputs: startDate, endDate, users.
get_plansB
Get plan mode adoption: which models are being used in plan mode and how often.
Inputs: startDate, endDate, users.
get_spendingB
Get current billing cycle spending for all team members. Shows spend in dollars, included vs overage, fast premium requests, and spend limits.
Inputs: page, allPages.
get_tabsB
Get tab autocomplete usage: suggestions shown vs accepted vs rejected, with line-level detail.
Inputs: startDate, endDate, users.
get_team_membersA
List all team members with their roles and status. Returns name, email, role, and whether they've been removed.
Inputs: no named parameters.
get_team_overviewA
Get a comprehensive team overview: member count, total spend, top spenders, DAU, and most-used models. This is the best starting point for understanding your team's Cursor usage.
Inputs: startDate, endDate.
get_usage_eventsA
Get granular per-request usage events with model, token counts, costs, and whether the request was chargeable. Supports filtering by user email and date range.
Inputs: email, startDate, endDate, page, +1 more.
get_user_deep_diveA
Deep dive into a specific user's usage: their spending, daily usage patterns, recent requests, and model preferences.
Inputs: email (required), startDate.
set_spend_limitA
Set a hard spending limit (in dollars) for a specific team member. Use with caution — this will block the user from making requests once the limit is reached.
Inputs: email (required), limitDollars (required).
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
TDQS
Scored across 17 tools
Every tool has a clearly distinct purpose targeting specific metrics or actions, such as get_agent_edits for edit metrics, get_daily_usage for user-level data, and set_spend_limit for administrative control. There is no overlap in functionality, making it easy for an agent to select the right tool without confusion.
All tool names follow a consistent verb_noun pattern with 'get_' prefix for retrieval operations and 'set_' for configuration, using snake_case throughout. This predictability enhances readability and usability for agents.
With 17 tools, the count is slightly high but reasonable for the comprehensive analytics and management scope of Cursor usage. It covers various aspects like billing, usage, and team management without feeling overly bloated.
The tool set provides complete coverage for analytics and administrative tasks in the Cursor domain, including metrics retrieval (e.g., usage, spending, models), team management, and spend control. There are no obvious gaps, enabling agents to handle all relevant workflows effectively.
Maintenance
Related MCP Connectors
Errors, logs and AI-call costs for apps on Node.js, Next.js, Cloudflare Workers or the browser. Ask your AI what’s broken, why, and how to fix it, right inside Cursor or Claude.
Track API spend, credit balances and runway across OpenAI, Anthropic, OpenRouter and 25 more.
Read-only analytics for Convex apps, queryable via MCP from Claude, Cursor, and other clients.
AI coding plan limits (Claude Code, Codex, Cursor, Copilot): router, picker, status, budget check
Related MCP Servers
- FlicenseBqualityNot gradedmaintenanceProvides comprehensive telemetry and usage analytics for Claude Code sessions, including token usage tracking, cost monitoring, and tool usage patterns. Enables users to monitor their Claude usage with detailed metrics, warnings, and trend analysis.12-
- FlicenseNot gradedqualityNot gradedmaintenanceProvides intelligent analysis of token usage patterns and optimization recommendations to improve efficiency and reduce costs in Claude Code sessions. Offers real-time analysis, cost metrics, and actionable insights for better context window and tool usage optimization.3 npm-
- AlicenseNot gradedqualityDmaintenanceTracks token usage and records coding sessions for Claude Desktop users, supporting both local SQLite and cloud Cloudflare D1 databases.556 npmMIT
- AlicenseAqualityCmaintenanceReal-time Claude.ai subscription awareness for AI coding assistants. Surfaces live utilization, forecasts limits, gates expensive operations, and measures real per-task cost.516 npm6MIT