What Is Trigger.dev?

Trigger.dev is an open-source background job and durable execution platform for TypeScript applications that lets developers run long-lived tasks, AI pipelines, and scheduled work without losing progress. This explainer shows what Trigger.dev does, how the run engine works, and how the repository is organized around execution reliability.

Trigger.dev at a Glance

Trigger.dev is an open-source background task and workflow platform for TypeScript. It's a large monorepo containing 5 applications, 8 published packages, and 14 internal packages that together deliver durable execution — tasks that survive server restarts, network failures, and long processing times.

The Product at a Glance

Core Concepts

Understanding Trigger.dev starts with its execution model. Everything revolves around tasks that run inside managed workers:

Organization
  ├── Projects
  │     ├── Environments (Development, Staging, Production, Preview)
  │     │     ├── Background Workers (versioned deployments)
  │     │     │     └── Tasks (the units of work you define)
  │     │     ├── Task Queues (concurrency + rate limits)
  │     │     ├── Task Schedules (cron-based triggers)
  │     │     └── Environment Variables
  │     └── Worker Groups (execution targets)
  ├── Members (team access)
  └── Access Tokens (API keys, PATs, org tokens)

The Tech Stack — Why Each Piece Exists

Key Statistics

Core Features

Task Execution

• No Timeouts — Tasks can run for minutes, hours, or days without hitting platform limits
• Automatic Retries — Configurable retry policies with exponential backoff
• Concurrency Control — Per-queue concurrency limits and rate limiting
• Idempotency — Built-in idempotency keys prevent duplicate execution
• Batch Triggering — Trigger thousands of tasks in a single API call

Durable Workflows

• Checkpointing — Save task state and restore after crashes or restarts
• Waitpoints — Pause execution waiting for external events, human approval, or time delays
• Sub-tasks — Tasks can trigger and wait for other tasks (nested execution)
• Debouncing — Collapse rapid triggers into a single execution

Scheduling & Triggers

• Cron Schedules — Define recurring tasks with cron expressions and timezone support
• Event Triggers — Trigger tasks from API calls, webhooks, or other tasks
• Delayed Runs — Schedule tasks to run at a specific future time

Developer Experience

• TypeScript-first SDK — Define tasks with full type safety and autocompletion
• Local Development — trigger dev runs tasks locally with hot reload via WebSocket
• CLI Deploy — trigger deploy builds and ships task code to the platform
• React Hooks — @trigger.dev/react-hooks for real-time run status in your UI
• AI Integration — toolFromTask bridges Trigger.dev tasks with the Vercel AI SDK

Observability

• Run Dashboard — Real-time monitoring of all task runs with status, logs, and traces
• OpenTelemetry — Full distributed tracing across SDK, workers, and platform
• ClickHouse Analytics — High-performance log and event querying
• TSQL — Custom query language for searching task logs
• Realtime Streaming — Live log output and run progress via Socket.io / SSE

Infrastructure

• Multi-tenant Isolation — Organization → Project → Environment hierarchy
• Environment Management — Dev, Staging, Production, and Preview environments with separate API keys
• Self-hosting — Docker Compose and Helm chart for full on-premise deployment
• Worker Groups — Route tasks to specific machine types or custom runtimes

What Makes Trigger.dev Different

1. Durable execution with checkpointing — Tasks survive crashes and restarts; no other open-source TypeScript platform offers this
2. TypeScript-native — Not a YAML-configured workflow engine; tasks are plain TypeScript functions
3. Human-in-the-loop — Waitpoints let tasks pause for external input (approvals, webhooks, manual data)
4. Custom runtimes — Run browsers (Playwright/Puppeteer), Python, FFmpeg, or any Docker image
5. Cloud + self-hosted — Same codebase powers the managed cloud and on-premise deployments
6. Fair multi-tenant queuing — MarQS (the internal queue system) ensures no single tenant starves others

What's Next

Resources

• Official Documentation
• GitHub Repository
• Cloud Platform
• Community Discord
• Self-hosting Guide