• EXECUTION
  • STOPPED
  • RETRY
  • HUMAN

How to manage task execution with your AI

Connect Trigger.dev to your agent to control background jobs. You can start new runs or fix old ones without leaving your chat.

Ask AI about this page

Short answer

How can I control my background tasks with AI?

You can trigger new jobs, stop running processes, or rerun failed tasks using your AI client. This gives you direct control over your backend logic. You'll see how these actions work below.

Task control outcomes

Where your jobs land.

Your agent handles the execution logic for you.

EXECUTION

New job starts

The AI kicks off a specific background process. You get a confirmation once the run is active.

STOPPED

Active runs end

The agent kills a specific process that is stuck or no longer needed. This stops unnecessary resource use.

RETRY

Failed runs fixed

Your AI re-runs a specific execution that didn't finish correctly. This helps clear errors automatically.

HUMAN

Manual control

The AI cannot change your code or delete your account. You keep full ownership of your infrastructure.

The workflow

What your AI does when <the thing arrives>.

The agent acts as the interface for your background workers.

  1. Start a job

    The agent identifies the task you want to run and initiates the process.

    trigger_task

  2. Kill a process

    If a run is hanging, the agent targets that specific ID to stop it.

    cancel_run

  3. Rerun a task

    The agent takes a previous run and starts it again to resolve failures.

    replay_run

  4. Manage execution

    The agent uses the available tools to keep your background tasks in sync.

    trigger_taskcancel_runreplay_run

Try it

Copy these to start.

Use these prompts to talk to Trigger.dev.

Starting points

Replace the task names and run IDs with your actual data.

Trigger.dev Connector

You're all set. Choose your MCP client and follow the setup instructions.

Connector linkhttps://edge.vinkius.com/vk_preview_E3D1PSFbED6190OhQ3gSMc4dXdMknoYFUrVm1Mgk/mcp

Claude Desktop

Follow the steps below to connect in seconds.

  1. 1In Claude Desktop, open Settings → Connectors.
  2. 2Click “Add custom connector” and paste the connector link above as the remote MCP server URL.
  3. 3Click Add and start a new chat — Trigger.dev capabilities are ready to use.
Configuration · claude_desktop_config.jsonCopy
{
  "mcpServers": {
    "triggerdev-mcp": {
      "url": "https://edge.vinkius.com/vk_preview_E3D1PSFbED6190OhQ3gSMc4dXdMknoYFUrVm1Mgk/mcp"
    }
  }
}
Copy into chat04

  • Trigger the 'sync-user-data' task for user_99.

  • Cancel the run with ID run_abc123.

  • Replay the failed 'process-payment' task.

  • Start a new run for the 'cleanup-logs' task.

  • Claude
  • ChatGPT
  • Cursor
  • VS Code
  • Windsurf
  • Claude Code
  • JetBrains
  • Cline

Start here

Connect Trigger.dev once, then ask.

Connect your account in one click. Your login stays encrypted on our side, and you can find the full tool list on the Connector page.

Connect Trigger.dev to your AI

FAQ

How this task behaves

  • 01

    Can the AI delete my tasks?

    No. The AI can only trigger, cancel, or replay existing tasks.

  • 02

    Can the AI edit my task code?

    No. The AI only executes the logic you have already written in your code.

  • 03

    How do I stop a specific task?

    Tell your agent to cancel the specific run ID you want to stop.

  • 04

    Can the AI create new task types?

    No. The AI can only interact with tasks that already exist in your Trigger.dev account.

  • 05

    What happens if a replay fails?

    The agent will report the failure, and you can decide whether to try again or investigate manually.

  • More questions about Trigger.dev? The Connector page answers them. See everything the Trigger.dev Connector can do

Connect Trigger.dev to Claude, Cursor, ChatGPT & more