Trigger.dev
Also known as: Trigger.dev v4, triggerdotdev
Open source TypeScript platform for long-running background tasks and AI agents with no timeouts, a chat agent runtime with stored conversations and tool approvals, on managed cloud or self-hosted under Apache 2.0.
Trigger.dev is an open source platform for writing long-running background tasks and AI agents in TypeScript. Tasks run with no timeouts on managed workers, trigger subtasks and batches and wait on them, and carry retries, queues, concurrency limits, idempotency, durable waits and versioned deploys. Scheduled tasks run on cron, and waits are checkpointed so idle time is not billed. Its AI agent guides cover prompt chaining, routing, parallelization, orchestrator-worker and evaluator-optimizer patterns.
The AI Chat runtime runs chat agents built on the Vercel AI SDK. Each conversation is stored as a transcript keyed by chat ID, so a new run picks up where the last one left off, with undo, edit and regenerate, and an adapter can keep transcripts in the customer's own database. Tools can be marked as needing a person's approval before they run.
AI Prompts version prompt templates on each deploy and let a team change a prompt's text or model from the dashboard without redeploying, and a testing harness runs an agent's turns against a mocked model. Every run is traced, with model calls recorded as spans carrying tokens, cost and latency. An official MCP server lets coding assistants start and message agents and manage runs.
Trigger.dev is Apache 2.0 licensed and self-hosts on the customer's own infrastructure with a setup guide, alongside its managed cloud. The Free plan costs $0 a month with $5 of usage credits, and Hobby at $10 and Pro at $50 a month include matching credits, billed on seconds of compute by machine size plus $0.25 per 10,000 runs. Enterprise adds role-based access control, single sign-on, and SOC 2 and penetration test reports, and a HIPAA BAA add-on is available from the Hobby plan.
Vendor details
Canonical URL
https://trigger.dev
Category
Agent infrastructure
Company status
independent
Use cases & customers
Target customers
Deployment options
In practice
Your agent run legitimately takes forty minutes and every serverless platform kills it. Trigger.dev lets you set a task's maximum duration or turn the limit off, and it checkpoints the task's full state during waits, which do not count toward that time.
Your compliance team will not accept a third party running the workload. Trigger.dev is open source under Apache 2.0 and can be self hosted on your own infrastructure, and Enterprise adds SOC 2 and penetration test reports, single sign on and role based access control, with a HIPAA BAA available as an extra.
The agent needs to drive a headless browser or run FFmpeg mid task. Tasks can install system packages and headless browsers through build extensions, so those tools run inside the task itself.
Sources & related URLs
Agentic Index coverage score
11.0 / 14 capabilities · 79%
| Integrations & Tool Calling | Partial |
|---|---|
|
The customer's agent code and its tools run on Trigger.dev. A chat.agent calls streamText with tools the customer defines through the Vercel AI SDK, and any task can call an outside API through its SDK or fetch. There is no connector catalog or authenticated action framework of Trigger.dev's own, since the tools and credentials are the customer's code. Sourcetrigger.dev/docs/guides/ai-agents/overviewread 2026-09-23 |
|
| Workflow Orchestration | Full |
|
Trigger.dev is a workflow engine for long-running tasks. Tasks trigger subtasks and batches and wait on them, with retries, queues, concurrency limits, idempotency, durable waits and versioned deploys. Its AI agent guides cover prompt chaining, routing, parallelization, orchestrator-worker and evaluator-optimizer patterns, and chat agents run sub-agents. Sourcetrigger.dev/docs/guides/ai-agents/overviewread 2026-09-23 |
|
| Knowledge Grounding & RAG | Not documented |
|
There is no retrieval structure that Trigger.dev maintains over the customer's corpus. Its example projects build retrieval and chat over the customer's data in the customer's own code, and the chat transcript store holds conversations. Sourcetrigger.dev/docs/guides/ai-agents/overviewread 2026-09-23 |
|
| Human Oversight & Guardrails | Full |
|
Agent actions can be held for approval in the AI Chat runtime. Tools flagged as needing approval pause the agent's tool call until a person answers, the runtime carries that pause across idle gaps, redeploys and crashes, and its human-in-the-loop pattern gates new actions while a tool is waiting on the answer. HITL approvals are among what the runtime handles for the customer. The customer sets that hold per tool, and it applies to the agent's own action, not just a general durable wait. Sourcetrigger.dev/docs/ai-chat/patterns/human-in-the-loopread 2026-09-23 |
|
| Security, Identity & Governance | Full |
|
Single sign on and role based access control come with Enterprise, and Pro adds AWS PrivateLink. Enterprise also includes a SOC 2 report and a penetration test report, and a HIPAA BAA is available as an extra from the Hobby plan up. Sourcetrigger.dev/pricingread 2026-09-23 |
|
| Observability & Auditability | Full |
|
Every run is recorded as a trace with logs and spans, and LLM observability turns each Vercel AI SDK call inside a task into a span carrying the model, provider, input, output, token counts, cost and latency, linked to prompt versions and queryable across runs. Chat sessions render in the dashboard's Sessions view, alerts go to up to 100 or more destinations, and log retention runs by plan, from 1 day on Free and 7 on Hobby to 30 on Pro and custom on Enterprise. Sourcetrigger.dev/docs/ai/observabilityread 2026-09-23 |
|
| Memory & State Persistence | Full |
|
A conversation outlives a single run. The AI Chat runtime keeps each chat.agent conversation in a transcript storage keyed by chatId, and when a run ends, whether through a version upgrade, its turn limit or a crash, the next message boots a fresh run that reads the conversation back to answer in context. The runtime writes the transcript from the agent's own turns, including tool calls and results, keyed per conversation. Sessions carry an expiresAt hard retention deadline, and streams age out on the plan's retention window. The transcript is read back through the loadTranscript and chat.history APIs and the dashboard's Sessions view, and changed through undo, edit and regenerate actions or a TranscriptStorage adapter to the customer's own database. Sourcetrigger.dev/docs/ai-chat/transcript-storageread 2026-09-23 |
|
| Deployment & Data Residency | Full |
|
Trigger.dev is open source under Apache 2.0 and self-hosts in the customer's own infrastructure following a self-hosting guide, alongside the managed cloud, where runs can state a region preference. Sourcetrigger.dev/docs/self-hosting/overviewread 2026-09-23 |
|
| Prebuilt Agents, Templates & Packs | Partial |
|
Ready made starting points ship with the product, including agent pattern guides for routing, translate and refine, and content checks, dozens of example projects, such as a chat agent over ClickHouse, a Cursor background agent, a batch LLM evaluator, human-in-the-loop workflows and Mastra agents with memory, and building-with-AI resources for coding agents. These are examples a developer builds from, not a catalog of agents a customer adopts. Sourcetrigger.dev/docs/guides/ai-agents/overviewread 2026-09-23 |
|
| Triggers & Channel Coverage | Full |
|
Scheduled tasks run on cron schedules declared in code or created from the dashboard or API, from 10 on Free up to 1,000 or more on Pro, and a chat agent wakes when a new message arrives on its session. Sourcetrigger.dev/docs/tasks/scheduledread 2026-09-23 |
|
| Model Flexibility & Routing | Full |
|
The customer chooses the model through Trigger.dev's own AI Prompts. Prompt templates carry a default model in code and are versioned on each deploy, and the team can create overrides from the dashboard that change the prompt text or the model without redeploying, with token usage, cost and latency tracked per prompt version. Chat agents resolve the named model through a provider registry. Sourcetrigger.dev/docs/ai/promptsread 2026-09-23 |
|
| APIs, SDKs & MCP Extensibility | Full |
|
Trigger.dev is callable from outside through its TypeScript SDK and management API for triggering and inspecting runs, the CLI, and an official MCP server whose tools let coding assistants list, start and message chat agents and manage runs. Sourcetrigger.dev/docs/ai-chat/mcpread 2026-09-23 |
|
| Testing, Debugging & Optimization | Partial |
|
Testing covers the agent's code. The AI Chat testing harness (mockChatAgent) runs the agent's real turn loop, lifecycle hooks and actions in memory with the model mocked, so a test asserts on chunks whose content the test itself scripted. AI Prompts track token usage, cost and latency per prompt version, so a team can compare versions on spend and speed. No quality score per prompt version is named. Sourcetrigger.dev/docs/ai-chat/testingread 2026-09-25 |
|
| Browser & Computer Use | Partial |
|
Browser workloads run in an environment Trigger.dev hosts. Tasks can install system packages and headless browsers through build extensions, and an example project drives a hosted browser from Anchor Browser inside a task. The control of the browser comes from those libraries and services in the customer's own code, and Trigger.dev has no API for screenshots, clicks or typing. Sourcetrigger.dev/docs/guides/example-projects/anchor-browser-web-scraperread 2026-09-23 |
|
The Agentic Index coverage score grades every vendor Full, Partial or Not documented against the same 14 buyer facing capabilities, from public evidence only. Each capability links to how all vendors in the index score on it. How this evidence is graded
Product updates
Inbound webhooks are now in public beta on Trigger.dev Cloud and the open source project, so events from Slack, Stripe, GitHub, Linear and more than 40 other services can start agents directly. Each webhook is verified, checked for duplicates and stored before it is routed to as many as 25 agents whose filters match. The dashboard shows every payload, filter, failure and replay alongside the steps of the agent it started.
Bears on: Workflow orchestration
View sourceAsk Trigger, now in beta, is an AI agent inside the Trigger.dev dashboard that can see a team's projects, environments, runs, errors, queues, tasks and deploys. It writes TRQL queries against Trigger.dev's analytics tables to explain failing or slow runs and build charts, and with a connected GitHub account it inspects the deployed source code and suggests a fix.
Bears on: Observability / auditability
View sourceTrigger.dev v4.7.0 moves concurrency limits from queues to tasks, so a per customer cap can also carry a ceiling across all customers, and named limits can be shared by several tasks and changed at run time. Chat agents can also cap concurrency per session.
Bears on: Workflow orchestration
View sourcePricing
Free at $0 a month with $5 in credits, Hobby at $10 a month and Pro at $50 a month, each plus usage.
A plan fee with matching monthly usage credits, then seconds of compute at a per machine rate plus a per run charge. The open source release is free to self host.
What is public
Trigger.dev publishes every plan, its limits, the per second compute rates by machine and the per run charge at trigger.dev/pricing, with a worked cost example. Enterprise is through sales.
Billing mechanics
Four managed cloud plans: Free, Hobby, Pro and Enterprise. Each paid plan's fee includes the same amount in monthly usage credits. Usage is billed on seconds of compute at a per machine rate, from Micro (0.25 vCPU, $0.0000169 per second) to Large 2x (8 vCPU, $0.00068 per second), plus $0.25 per 10,000 runs. Self hosting the Apache 2.0 release carries no vendor charge.
Cost watchouts
Every run that starts on the managed workers adds an invocation charge on top of compute, and the machine size sets the per second rate. On Pro, extra concurrency, seats, preview branches, dashboards and schedules are sold in bundles at $10 to $20 a month each. RBAC, SSO and the SOC 2 report are Enterprise only.
Variable cost rationale
Trigger.dev meters seconds of compute at a rate set by machine size, from $0.0000169 per second on Micro to $0.00068 on Large 2x, plus $0.25 per 10,000 runs. Run length, machine choice and run volume make the bill grow, and a long agent task keeps billing compute until it ends. Each paid plan's fee comes back as the same amount in monthly usage credits, and only usage beyond them is billed. A task's maximum duration caps a run, though it can be raised or turned off per task. Checkpointed waits and parent tasks waiting on subtasks are not billed, runs in the DEV environment are free, and self hosting the Apache 2.0 release carries no vendor charge.
Additional watchouts
Governance features (RBAC, SSO, the SOC 2 and penetration test reports) cannot be bought self serve. All plans are limited to 1,500 API requests per minute, which can be raised on paid plans.
Overage / add-ons
Usage beyond the plan's included credits is billed per second of compute at the machine's rate plus $0.000025 per run. Runs in the DEV environment are free.
Sales call required
Mixed (some tiers require a call)
Free / trial
Free plan at $0 a month with $5 of monthly usage credits, and the open source release is free to self host
Lowest paid plan
Hobby at $10 a month, with $10 of usage credits included
Commercial notes
Waiting is free. Checkpointed waits and parent tasks waiting on subtasks count toward neither compute nor concurrency, which suits long agent tasks that spend most of their time waiting on models, tools or people. Trigger.dev publishes a worked cost example beside its plans.
Key ambiguities
Monthly cost depends on machine size, run length and run count. Waits longer than five seconds and parent tasks waiting on subtasks are checkpointed and not billed. Enterprise pricing and the HIPAA BAA add-on price are through sales.
Support SLA / resale
Community support on Free and Hobby, dedicated Slack support on Pro, and priority support on Enterprise.
Missing data
Enterprise pricing and the HIPAA BAA add-on price.
Related vendors
- AgentOps — Agent observability and debugging platform: open source SDKs trace…
- AgentPhone — AgentPhone gives the AI agents a team builds their own US or…
- Agno — Python agent framework and AgentOS runtime (formerly Phidata) for…
- AIsa — Resource and payment gateway for AI agents, with one key to 110+…
- Albato — Albato is a no code integration platform that sells an embedded,…
- AlphaBitCore — AI control plane for regulated financial firms: one gateway enforces…
Alternatives to Trigger.dev
The closest documented capability profiles to Trigger.dev among agent infrastructure platforms tracked by Agentic Index, ordered by similarity on the same 14 point evidence the rankings use. No vendor pays for placement.
- Netlify10.5 / 14Adds documented Knowledge Grounding & RAG
- Inngest10.0 / 14Fuller documented coverage on Testing, Debugging & Optimization
- Pipedream11.0 / 14Fuller documented coverage on Integrations & Tool Calling and Browser & Computer Use
- Thread AI11.0 / 14Adds documented Knowledge Grounding & RAG
- Windmill12.0 / 14Fuller documented coverage on Integrations & Tool Calling and Prebuilt Agents, Templates & Packs
- xpander.ai12.0 / 14Adds documented Knowledge Grounding & RAG
Similarity is computed from each vendor's Agentic Index coverage score evidence, axis by axis, not from the totals. How this evidence is graded