Trigger.dev review

Trigger.dev is an open-source TypeScript platform for running durable background tasks and workflows. In the hosted setup, your application defines tasks and sends work to Trigger.dev, which runs that work on its infrastructure. The platform handles queues, retries, long-running execution, scaling and run observability. The vendor also offers self-hosting, so it can sit on your own infrastructure rather than only as a hosted service.

It fits between an application and the work that should not run inside a short-lived request: scheduled jobs, webhook-triggered tasks, media processing and multi-stage workflows. For AI applications, tasks can call tools, stream responses and pause for human approval. The homepage says tasks are written in a Trigger folder and describes atomic deploy versions, so started tasks are not affected by code changes. It is workflow infrastructure, not a model provider or a general-purpose queue library. Teams should assess its task model and operational needs against the job system already in their stack.

The homepage presents Trigger.dev as a TypeScript platform where you write tasks as code in a Trigger folder. Its AI-agent example uses TypeScript, chat.agent, typed tools and run; the example also integrates Anthropic through anthropic("claude-sonnet-4-5") and streaming through streamText. Those are example integrations, not a requirement to use that model provider.

The documented integration surfaces named on the homepage include webhooks, a Realtime API for task status and streams, React hooks, an MCP Server and Python support. Python scripts can run with automatic package installation through requirements.txt; the main platform is described as being for TypeScript. Build customization includes esbuild plugins, Prisma, Puppeteer, FFmpeg, apt-get and additional packages. The homepage also lists durable schedules, waits, batch triggering and typed, validated payloads. It does not name a CLI or enumerate SDK packages, so confirm SDK and CLI support, authentication details and current API contracts in the vendor's docs before choosing an integration.

The Free plan costs $0 per month and includes $5 in monthly compute credits, 20 concurrent runs, unlimited tasks, five team members and 10 schedules. Hobby costs $10 per month, with $10 in monthly credits, 50 concurrent runs, five team members, 100 schedules and seven-day log retention. Pro starts at $50 per month and includes $50 in credits, 200+ concurrent runs, 25+ team members, 1,000+ schedules and 30-day log retention. Enterprise pricing is custom.

The bill can grow beyond the plan price. The pricing page lists compute at $0.0000169–$0.0006800 per second, depending on machine, and run invocations at $0.000025 each. Pro add-ons include $10 per additional 50 concurrent runs above 200, $20 per seat above 25, and $10 per 1,000 schedules above 1,000. Extra preview branches, dashboards and Realtime connections are also $10 each at their listed thresholds. A HIPAA BAA is $500 per month. Prices are monthly; annual billing or discounts are not published.

Trigger.dev describes itself as open source, and its community post points to the project’s GitHub repository and a Docker self-hosting tutorial. The vendor’s homepage also says self-hosting is an option. The licence name and any licence-specific restrictions are not stated on the homepage or pricing details, so check the repository’s current licence before using or modifying the code in a commercial deployment.

Self-hosting gives teams an alternative to relying entirely on the hosted service, but it also means operating the infrastructure. The vendor does not publish a data-export path in the materials reviewed for this page; ask how to export run history, task state and relevant metadata before adoption. Workflow code is TypeScript, which can reduce the cost of moving task logic, but Trigger-specific task APIs, retries and run state may require rewriting or a migration plan. Treat portability of in-flight runs as a separate question from portability of source code.

The vendor describes elastic scaling, retries, queues, alerts, advanced run filtering and atomic deploy versions. Its homepage also promotes multi-region workers, static IPs and AWS PrivateLink. The Pro plan lists 200+ concurrent runs, 1,000+ schedules and 30-day log retention; check the exact limits and any add-on costs for your tier before sizing a production workload. The homepage’s “no timeouts” claim should be validated against the task types and operational guarantees your application needs.

For security and compliance, Enterprise lists SSO and RBAC, while the pricing page lists a HIPAA BAA for $500 per month. Confirm data handling, retention and recovery commitments with the vendor for your workload. Reddit feedback is sparse but includes a TypeScript user praising setup and a commenter noting that Trigger.dev charges by job invocation rather than step. Another Reddit thread raised questions about how the system works, a reminder to verify its execution model and deployment fit rather than treating it as a drop-in queue. The vendor does not publish a service-level commitment in the details reviewed here.