Cursor

Build workflows with Cursor AI Agents and track usage

Actions

Get Daily Usage Data Fetches daily team usage metrics from Cursor's Admin API.

Get Last Message Retrieves the last message from a Cursor Cloud Agent conversation.

Launch Cloud Agent Launches a Cursor Cloud Agent to perform coding tasks asynchronously.

Instructions

To get your API keys, visit the Cursor Dashboard. You may need separate keys for Agents and Admin features.

Get Daily Usage Data

The Get Daily Usage Data component fetches team usage metrics from Cursor’s Admin API.

Use Cases

• Usage reporting: Track team productivity and AI usage patterns
• Cost tracking: Monitor usage-based requests and subscription consumption
• Analytics dashboards: Build custom dashboards with Cursor usage data

How It Works

1. Fetches usage data for the specified date range from Cursor’s Admin API
2. Returns detailed metrics per user including lines added/deleted, requests, and model usage

Configuration

• Start Date: Start of the date range (YYYY-MM-DD format, defaults to 7 days ago)
• End Date: End of the date range (YYYY-MM-DD format, defaults to today)

Output

The output includes per-user daily metrics:

• Lines added/deleted (total and accepted)
• Tab completions shown/accepted
• Composer, chat, and agent requests
• Subscription vs usage-based request counts
• Most used model and file extensions

Notes

• Requires a valid Cursor Admin API key configured in the integration
• Only returns data for active users

Example Output

{
  "data": [
    {
      "acceptedLinesAdded": 1102,
      "acceptedLinesDeleted": 645,
      "agentRequests": 12,
      "chatRequests": 128,
      "composerRequests": 45,
      "date": 1710720000000,
      "email": "developer@company.com",
      "isActive": true,
      "mostUsedModel": "gpt-4",
      "subscriptionIncludedReqs": 180,
      "totalAccepts": 73,
      "totalApplies": 87,
      "totalLinesAdded": 1543,
      "totalLinesDeleted": 892,
      "totalRejects": 14,
      "totalTabsAccepted": 289,
      "totalTabsShown": 342,
      "usageBasedReqs": 5
    }
  ],
  "period": {
    "endDate": 1710892800000,
    "startDate": 1710720000000
  }
}

Get Last Message

The Get Last Message component retrieves the last message from a Cursor Cloud Agent’s conversation history.

Use Cases

• Message tracking: Get the latest response or prompt from an agent conversation
• Workflow automation: Use the last message as input for downstream components
• Status monitoring: Check what the agent last communicated

How It Works

1. Fetches the conversation history for the specified agent ID
2. Extracts the last message from the conversation
3. Returns the message details including ID, type (user_message or assistant_message), and text

Configuration

• Agent ID: The unique identifier for the cloud agent (e.g., bc_abc123)

Output

The output includes:

• Agent ID: The identifier of the agent
• Message: The last message object containing:
  • ID: Unique message identifier
  • Type: Either “user_message” or “assistant_message”
  • Text: The message content

Notes

• Requires a valid Cursor Cloud Agent API key configured in the integration
• If the agent has been deleted, the conversation cannot be accessed
• Returns nil if the conversation has no messages

Example Output

{
  "agentId": "bc_abc123",
  "message": {
    "id": "msg_005",
    "text": "I've added a troubleshooting section to the README.",
    "type": "assistant_message"
  }
}

Launch Cloud Agent

The Launch Cloud Agent component triggers a Cursor AI coding agent and waits for it to complete.

Use Cases

• Automated code generation: Generate code from natural language prompts
• PR fixes: Automatically fix issues on existing pull requests
• Code refactoring: Refactor code based on instructions
• Feature implementation: Implement new features from specifications

How It Works

1. Launches a Cursor Cloud Agent with the specified prompt and configuration
2. Waits for the agent to complete (monitored via webhook and polling)
3. Emits output with the agent result (success or failure)

Example Output

{
  "agentId": "agent_12345",
  "branchName": "cursor/agent-550e8400",
  "prUrl": "https://github.com/org/repo/pull/42",
  "status": "done",
  "summary": "Refactored login logic."
}