# Trigger.dev

Build and deploy durable AI agents and workflows

Source: https://wheredoirunthis.com/products/trigger-dev · JSON: https://wheredoirunthis.com/api/products/trigger-dev · Read on 2026-10-15

## At a glance

- Type: Workflow engine
- Kind: Open core
- Isolation: Container
- Free tier: Yes
- Starting price: $10/mo
- GPU: No
- Regions: Not disclosed
- Coverage: 18 of 61 assessed capabilities supported
- Agent readiness: read 10.5 of 11 signals, drive 10.5 of 11 signals

## Links

- Homepage: https://trigger.dev/
- Docs: https://trigger.dev/docs
- Pricing: https://trigger.dev/pricing
- Repository: https://github.com/triggerdotdev/trigger.dev

## Pricing

- Model: plan tier (quotas + features) + usage (per-second compute per machine size + per-run invocation fee)
- Unit: compute-second (per machine size) + per-run invocation ($0.25/10,000 runs); paused/checkpointed time NOT billed
- Free tier: Yes

| Tier | Price | Includes | Notes |
| --- | --- | --- | --- |
| Free | $0/mo | $5/mo free credits · 20 concurrent runs · unlimited tasks · 5 team members · 10 schedules · 1-day logs/query · 10 Realtime conns · 1 alert destination · preview branches; $5 usage credit/mo (dollar-denominated); 20 concurrent (prod), 10 staging/10 preview/25 dev (limits docs); 5 team members; DEV-environment runs free |  [source](https://trigger.dev/pricing) |
| Hobby | $10/mo | $10/mo credits · 50 concurrent · 100 schedules · 7-day logs · 5 preview branches · 3 alert destinations · 150 Realtime conns · HIPAA BAA add-on; $10 usage credit; 5 preview branches; HIPAA BAA requestable as paid add-on |  [source](https://trigger.dev/pricing) |
| Pro | $50/mo | $50/mo credits · 200+ concurrent ($10/50) · 25+ members ($20/seat) · 20+ preview branches ($10/ea) · 1000+ schedules ($10/1k) · 30-day logs/query · 1000+ Realtime conns ($10/1k) · 5+ custom dashboards ($10/ea) · AWS PrivateLink · HIPAA BAA add-on · dedicated Slack support; $50 usage credit; seats beyond 25 and concurrency beyond 200 purchased in blocks; AWS PrivateLink included on Pro |  [source](https://trigger.dev/pricing) |
| Enterprise | custom | All Pro + custom discounts, custom log retention, RBAC, SOC 2 report, penetration test report, SSO, priority support, HIPAA BAA | Contact us — not estimable from the public surface; SOC 2 report + pen-test report are Enterprise-tier deliverables (the SOC 2 Type II certification itself is company-held, per /security). [source](https://trigger.dev/pricing) |

- Overages: per-run $0.000025 ($0.25/10k); compute per-second per machine (Small 1x $0.0000338/s ≈ $0.122/hr → Large 2x 8vCPU/16GB $0.00068/s); add-on blocks: +50 concurrency $10, +seat $20, +preview branch $10, +dashboard $10, +1k schedules $10, +1k Realtime conns $10
- Card required: No
- Source: https://trigger.dev/pricing%20(fresh%20markdown%20capture%202026-10-15)

## Website design

Desktop public website · assessed 2026-10-15

Complete assessment.

Dark, restrained developer-tool marketing site: near-black surfaces (body background rgb(18,19,23)), soft off-white text, generous whitespace and a large type hierarchy, with one vivid lime-green accent (measured rgb(168,255,83) on CTAs and active nav) used sparingly. Pricing uses a clean four-column card grid with fine low-contrast rules; docs shares the dark theme with a conventional 304px sidebar, search and lime active-state markers. A blurred terminal hero image and a cookie-consent banner (landing/pricing) are visible; no broken images or failed renders on any of the three pages.

- Styles: Editorial minimalism
- Landing page: captured
- Pricing: captured
- Docs: captured

## Runtime

- **GPU & accelerators**: No — "Machine presets top out at Large 2x (8 vCPU / 16 GB); no GPU machines. GPU/LLM inference is called out to providers from task code." [source](https://trigger.dev/pricing) (confidence: high)
- **BYOC (Bring your own cloud)**: advertised (placeholder) — "/byo-cloud still serves the /pricing content verbatim (byte-identical markdown, 'Source: https://trigger.dev/pricing') — a placeholder route, unchanged since the 2026-05 intel. No real BYOC flow confirmed on the public surface." [source](https://trigger.dev/byo-cloud) (confidence: low)
- **Secrets management**: Yes — "Environment variables/secrets per environment; every preview branch gets its own TRIGGER_SECRET_KEY and env vars." [source](https://trigger.dev/docs) (confidence: medium)
- **Storage options**: ephemeral per-run; durable run state via managed checkpoint storage (Cloud); self-host operators provide object storage themselves — "Logs in ClickHouse (queries via TRQL); checkpoints stored on Cloud-managed storage — a self-host dep alongside object storage." [source](https://trigger.dev/docs/self-hosting/overview) (confidence: high)
- **Supported workloads**: ai-agents, ai-chat-agents, background-jobs, scheduled-tasks, etl/rags, media-processing, human-in-the-loop, browser-automation — "llms.txt: 'Primary use cases: AI agents, multi-step LLM workflows, and RAG pipelines; long-running background jobs that exceed serverless limits; video, audio, image, and document processing; scheduled tasks and managed cron; multi-tenant SaaS background jobs.' chat.agent adds durable AI chat agents." [source](https://trigger.dev/llms.txt) (confidence: high)
- **Persistence**: checkpoint-resume (no persistent-disk; runs ephemeral, state snapshotted at wait points) — "Checkpoint-resume snapshots: 'paused agents release compute, use no resources and aren't billed while waiting, and resume exactly where they left off.' Cloud runs have an enforced max TTL of 14 days queued (limits docs)." [source](https://trigger.dev/llms-full.txt) (confidence: high)
- **Data jurisdiction**: multi-region workers (region list behind the dashboard); subprocessors in USA (AWS, DigitalOcean, Cloudflare, ClickHouse) + EU provider present — "Sub-processor list: AWS (USA, cloud infrastructure), DigitalOcean (USA), Cloudflare, ClickHouse, Depot (CI/CD), Anthropic/OpenAI (gen AI), PlanetScale, PostHog (EU)… Multi-region task deployment documented; static IPs 'for each region listed on the Regions page in the dashboard' — no public region enumeration." [source](https://trigger.dev/legal/subprocessors) (confidence: medium)
- **Resource limits & quotas**: Concurrency per environment (prod: Free 20 / Hobby 50 / Pro 200+, then $10 per +50); API rate limit 1,500 req/min all tiers (higher on request; batchTrigger up to 1,000 tasks/call); queued-run TTL cap 14 days (Cloud); machines Micro 0.25 vCPU/0.25 GB → Large 2x 8 vCPU/16 GB; Realtime connections Free 10 → Pro 1000+ ($10/1k); schedules Free 10 → Pro 1000+ ($10/1k) [source](https://trigger.dev/docs/limits%20+%20https://trigger.dev/pricing)
- **SDK languages**: typescript, javascript, python (via build extension only) — "'Normal TypeScript, no DSL … any npm package'; llms-full Quick Facts: 'TypeScript/JavaScript (Node.js; experimental Bun runtime); Python via build extension'. @trigger.dev/sdk@latest on npm = 4.7.4 (2026-10-15)." [source](https://trigger.dev/llms.txt) (confidence: high)
- **Network & egress controls**: Not documented — "No documented outbound egress firewall/allowlist on the public surface. Inverse exists: static egress IPs for allowlisting (paid) and AWS PrivateLink (Pro) — inbound/connectivity conveniences, not egress policy." (source: docs + limits + pricing scanned 2026-10-15) (confidence: high)
- **Snapshots & forks**: Yes — "'Checkpoint-resume snapshots capture full process state during waits: paused agents release compute, use no resources and aren't billed while waiting, and resume exactly where they left off.'" [source](https://trigger.dev/llms.txt) (confidence: high)
- **Deployment methods**: cli, git-deploy [source](https://trigger.dev/docs%20(CLI%20deploy;%20preview%20branches%20'manual%20or%20automatic%20via%20GitHub%20Action'))
- **Isolation**: Container — "'When a task calls wait.for() … the container is snapshotted, the memory and process state are persisted, and the container is freed. When the wait resolves, the snapshot is restored and execution continues from the exact instruction after the wait.' Per-run container with configurable CPU/RAM (machine presets)." [source](https://trigger.dev/llms-full.txt) (confidence: high)

## Capabilities

### AI & agents

| Capability | Status | Evidence |
| --- | --- | --- |
| Agent skills | Supported | "Install Trigger.dev agent skills to teach any AI coding assistant how to write tasks, realtime frontends, and chat.agent AI agents … npx trigger.dev@latest skills" [source](https://trigger.dev/docs/skills) |
| AI app builder | Not supported | "No prompt-to-app generation surface; the product is a code-first task runtime." [source](https://trigger.dev/) |
| Autonomous QA agent | Not supported | "'Ask Trigger' (dashboard, 2026-10-05) answers questions about your project — no first-party agent QA-tests a running app for defects." [source](https://trigger.dev/changelog) |
| Built-in coding agent | Not supported | "chat.agent is infrastructure YOU use to build durable AI agents; 'Ask Trigger' (changelog, Oct 5) is a dashboard assistant about your runs — neither is a first-party coding agent." [source](https://trigger.dev/docs/ai-chat/overview) |
| BYOK for LLMs | Not applicable | "LLM keys are plain secrets/env vars used in your task code — not a platform BYOK feature (secrets-management is the platform surface)." [source](https://trigger.dev/docs) |
| Credential-injection proxy | Not supported | "Secrets are injected as plain env vars the workload holds; no documented on-the-wire credential substitution keeping real keys outside the run." [source](https://trigger.dev/docs) |
| LLM gateway | Not supported | "Works with the Vercel AI SDK and any LLM provider — you call LLMs directly from task code; no platform gateway/proxy." [source](https://trigger.dev/llms.txt) |
| MCP client | Not supported | "The docs describe exposing chat agents AS tools through Trigger.dev's MCP server; the platform is not documented as consuming external MCP servers." [source](https://trigger.dev/docs/ai-chat/mcp) |
| MCP-server hosting | Not supported | "Tasks are async functions, not addressable servers — you cannot host a long-lived MCP server as a workload (chat.agent agents are reachable only through Trigger.dev's own MCP server)." [source](https://trigger.dev/docs) |
| MCP server | Supported | "npx trigger.dev@latest install-mcp … The Trigger.dev MCP Server enables AI assistants to interact directly with your Trigger.dev projects" [source](https://trigger.dev/docs/mcp-introduction) |
| Model inference API (first-party) | Not supported | "You bring the model providers (Vercel AI SDK, any LLM key); Trigger.dev serves no first-party inference endpoint." [source](https://trigger.dev/llms.txt) |
| Preinstalled coding agent | Not supported | "The relationship runs the OTHER direction — Trigger.dev installs ITS skills into the user's agent tools; it does not provision workspaces with third-party coding agents preinstalled." [source](https://trigger.dev/docs/skills) |

### Networking & delivery

| Capability | Status | Evidence |
| --- | --- | --- |
| Automatic HTTPS | Not applicable | "Task runtime, not a web-service host; does not provision TLS endpoints per workload." [source](https://trigger.dev/docs) |
| Custom domains | Not applicable | "No user-facing service hosting → no custom domains." [source](https://trigger.dev/docs) |
| Direct public IPv6 | Not supported | "No per-instance public addressing; outbound is via per-region static egress IPs (a shared-egress model, not per-run addresses)." [source](https://trigger.dev/docs/limits) |
| Edge KV store | Not supported | "No PoP-colocated KV store; caching/state happens in your own stores." [source](https://trigger.dev/docs) |
| Egress controls | Not supported | "No documented outbound egress firewall/allowlist; the docs cover concurrency and rate limits only" [source](https://trigger.dev/docs/limits) |
| Email send / receive | Not applicable | "Email sequences are a workload you BUILD in task code, not a platform send/receive primitive." [source](https://trigger.dev/llms.txt) |
| Public preview URLs | Not applicable | "Preview BRANCHES exist (per-branch isolated environments), but there is no shareable per-deploy public URL primitive — run state is shared via Realtime subscriptions instead (sharing partial)." [source](https://trigger.dev/docs) |
| Private networking | Partially supported | "Pro plan: AWS PrivateLink — Connect to private AWS resources (RDS, ElastiCache, internal APIs) without exposing them publicly: traffic never touches the public internet (llms-full)." [source](https://trigger.dev/pricing) |
| Built-in reverse proxy | Not applicable | "Not an ingress/routing product; tasks are not addressable HTTP services." [source](https://trigger.dev/docs) |

### Organisation & security

| Capability | Status | Evidence |
| --- | --- | --- |
| Built-in auth | Not applicable | "Dashboard/CLI auth for the platform itself (oauth GitHub login); not auth-as-a-service for user workloads." [source](https://trigger.dev/docs) |
| BYOC | Partially supported | "/byo-cloud still serves the pricing page content verbatim (placeholder route, byte-identical markdown, re-verified 2026-10-15)" [source](https://trigger.dev/byo-cloud) |
| Deploy into customers' VPCs | Not supported | "No product for deploying Trigger.dev software into YOUR customers' clouds; closest surfaces are the BYOC placeholder and self-host." [source](https://trigger.dev/pricing) |
| Secrets management | Supported | "Environment variables/secrets per environment; every preview branch gets its own TRIGGER_SECRET_KEY and env vars." [source](https://trigger.dev/docs/deploy-environment-variables) |
| Self-host / OSS core | Supported | "Open source, Apache 2.0. The full stack (dashboard, orchestrator, workers) self-hosts via Docker Compose or Kubernetes." [source](https://trigger.dev/docs/self-hosting/overview) |
| Sharing | Partially supported | "Realtime run/token streams + public API key variants expose run state to frontend subscribers — a run-state surface, not a general share-a-resource-by-link primitive." [source](https://trigger.dev/docs/realtime/overview) |
| SSO / SAML | Partially supported | "Enterprise: Single sign-on (SSO)" [source](https://trigger.dev/pricing) |
| Teams | Supported | "Organizations: 5 team members (Free) → 25+ on Pro, then $20/seat" [source](https://trigger.dev/pricing) |
| Workload identity | Not supported | "No OIDC/SPIFFE/IAM-federation workload identity; tasks authenticate outward with stored secrets (TRIGGER_SECRET_KEY is the platform API key)." [source](https://trigger.dev/docs) |

### Developer access

| Capability | Status | Evidence |
| --- | --- | --- |
| CLI | Supported | "Official trigger.dev CLI (init / dev / deploy / login / skills / install-mcp; npm package trigger.dev)" [source](https://trigger.dev/docs/cli-introduction) |
| E2B-compatible API | Not supported | "No E2B API/SDK compatibility surface." [source](https://trigger.dev/docs) |
| Filesystem API | Not supported | "No remote file-operation API over the management API; files are handled inside task code." [source](https://trigger.dev/docs) |
| Git deploy | Supported | "GitHub repo integration; preview branches deploy 'Manual or automatic via GitHub Action' / docs cli-deploy-commands + github-repo docs" [source](https://trigger.dev/docs/github-integration) |
| HTTP data API | Not supported | "The management API is a control plane (triggering/managing runs); there is no data-plane HTTP query surface — it stores no user data." [source](https://trigger.dev/docs/management/overview) |
| OpenAPI spec | Supported | "openapi: 3.1.0 — info: title: Trigger.dev v3 REST API (248KB spec referenced from the docs Mintlify config; verified 200 on 2026-10-15)" [source](https://trigger.dev/docs/v3-openapi.yaml) |
| Private container registry | Not supported | "No hosted private image registry; custom images enter via build extensions/Docker layers in builds, not a platform registry." [source](https://trigger.dev/docs) |
| Public API | Supported | "The REST API lets you trigger and manage runs — plus an OpenAPI 3.1 spec (see openapi-spec)." [source](https://trigger.dev/docs/management/overview) |
| SDK | Supported | "@trigger.dev/sdk (4.7.4 current) + Realtime React hooks + chat.agent (AI SDK) — TypeScript-first; Python via build extension." [source](https://trigger.dev/docs) |
| SSH access | Not supported | "Managed task runtime — no SSH into workers (hence no derived might-work harnesses)." [source](https://trigger.dev/docs) |

### Compute & storage

| Capability | Status | Evidence |
| --- | --- | --- |
| Cloud-init / ignition config | Not supported | "Runs on plain Docker Compose/K8s/Helm — no cloud-init/user-data launch-metadata interface (deploy_interface is CLI/Git, image-set via pinned runtime images)." [source](https://trigger.dev/docs/self-hosting/overview) |
| Desktop / computer-use sandbox | Not supported | "Playwright/Puppeteer build extensions give headless browser automation INSIDE a task — no GUI virtual desktop / computer-use sandbox." [source](https://trigger.dev/docs/config/extensions/overview) |
| GPU compute | Not supported | "machine presets: Micro 0.25 vCPU … Large 2x 8 vCPU/16 GB — no GPU machines" [source](https://trigger.dev/pricing) |
| Live-state (memory) snapshots | Supported | "Checkpoint-resume snapshots capture memory, CPU registers, and file descriptors … resume exactly where they left off; llms-full: 'the memory and process state are persisted' — checkpoints restore LIVE PROCESSES, not a freshly booted container." [source](https://trigger.dev/llms.txt) |
| Multi-region | Supported | "Multi-region workers: Deploy tasks to multiple regions for lower latency to users or data sources. … static IP addresses for each region are listed on the Regions page in the dashboard." [source](https://trigger.dev/llms-full.txt) |
| Object storage | Not supported | "No managed S3-compatible object store offered to workloads; object storage appears only as a SELF-HOST dependency (source.self_host.deps), the opposite direction." [source](https://trigger.dev/docs) |
| Persistent disk | Not applicable | "runs are ephemeral per-run containers; durability is checkpoint-resume snapshots, not attached volume storage" [source](https://trigger.dev/llms-full.txt) |
| Sandbox templates / custom images | Partially supported | "Build Extensions: Prisma, FFmpeg, Playwright, Python, Puppeteer, custom Docker layers with one config line; self-hosters pin runtime images by digest (node-22/24/26, bun)" [source](https://trigger.dev/docs/config/extensions/overview) |
| Scale to zero | Supported | "Elastic scaling without workers: Concurrency scales from 0 to your plan limit automatically … Paused time uses no resources and isn't billed" [source](https://trigger.dev/llms.txt) |
| Selectable isolation runtime | Not supported | "Selectable 'runtimes' are LANGUAGE runtimes (node-22/24/26, experimental bun) pinned per deploy — a single container isolation substrate, no per-workload isolation choice." [source](https://trigger.dev/docs/self-hosting/env/webapp) |
| Snapshots / forking | Supported | "the container is snapshotted, the memory and process state are persisted … when the wait resolves, the snapshot is restored and execution continues from the exact instruction after the wait" [source](https://trigger.dev/llms-full.txt) |

### Data

| Capability | Status | Evidence |
| --- | --- | --- |
| Database branching | Not supported | "A workflow engine, not a data store — no copy-on-write DB branch primitive (branching applies to preview ENVIRONMENTS, not databases)." [source](https://trigger.dev/docs/guides) |
| Ephemeral instant DB (no-signup) | Not supported | "No no-signup instant provisioning; signup is required (Free tier, no card)." [source](https://trigger.dev/pricing) |
| First-party database | Not supported | "Docs teach connecting YOUR database from task code; the platform provisions no managed user DB (ClickHouse/Postgres are internal)." [source](https://trigger.dev/docs/database-connections) |
| Vector search | Not supported | "RAG pipelines are a workload you build against your own stores; no first-party embedding/vector index." [source](https://trigger.dev/llms.txt) |

### Platform & operations

| Capability | Status | Evidence |
| --- | --- | --- |
| Managed CI/CD runners | Not supported | "Depot appears only as the vendor's own CI/CD build sub-processor — Trigger.dev does not offer hosted CI runner VMs as a product." [source](https://trigger.dev/legal/subprocessors) |
| Built-in observability | Supported | "OpenTelemetry traces, LLM observability (model, token, cost, and latency spans per AI SDK call), TRQL (ClickHouse-backed query language), custom dashboards, alerting" [source](https://trigger.dev/product/observability-and-monitoring) |
| Preview environments | Supported | "Isolated environment per git branch under the Preview environment. Each branch gets its own TRIGGER_SECRET_KEY, env vars, schedules, Realtime, and run history … Auto-archived when the PR closes." [source](https://trigger.dev/llms-full.txt) |
| Scheduled tasks | Supported | "Managed cron, per-environment, multi-tenant dynamic schedules." [source](https://trigger.dev/product/scheduled-tasks) |
| Project template marketplace | Partially supported | "Example-projects docs gallery — cloneable full starter projects (ask-trigger-chat-agent, clickhouse-chat-agent, cursor-background-agent, mastra-agents-with-memory, …)" [source](https://trigger.dev/docs/guides/example-projects/ask-trigger-chat-agent) |
| Durable workflows | Supported | "Durable execution via checkpoint-resume — THE core primitive: task() with retries, queues, waits, idempotency keys, at-least-once delivery (llms-full Quick Facts)." [source](https://trigger.dev/llms.txt) |

## Agent readiness

### Read: 10.5 of 11 signals present

10.5/11 signals (10 full, 1 partial)

- Llms txt: pass — "GET /llms.txt 200, 7.7KB, last updated 2026-10-09 — links docs + pricing + product pages + comparisons + a for-ai-crawlers index (https://trigger.dev/for-ai-crawlers)"
- Llms full txt: pass — "GET /llms-full.txt 200, 28KB 'Extended technical documentation' (updated 2026-10-08) with Quick Facts table + full exposition — CHANGED from fail (2026-06: not located)"
- Robots AI policy: pass — "robots.txt (946B) explicitly Allows GPTBot, ChatGPT-User, OAI-SearchBot, Claude-User/Chat/Web/Bot, PerplexityBot, Cohere-ai, Google-Extended, Applebot-Extended, Meta-ExternalAgent, anthropic-ai"
- JSON LD: pass — "homepage carries application/ld+json SoftwareApplication: name 'Trigger.dev', applicationCategory DeveloperApplication, description, url"
- Sitemap: pass — "robots.txt Sitemap pointers → GET /sitemap.xml 200, 22KB, non-trivial (also blog/changelog/for-ai-crawlers RSS)"
- Meta quality: pass — "homepage: title 'Trigger.dev | The open source platform for durable AI agents' + og:title + descriptive meta description"
- Markdown negotiation: pass — "VERIFIED this scan: curl -H 'Accept: text/markdown' https://trigger.dev/pricing returns text/markdown (15.6KB); llms.txt also documents 'append .md to the URL … or request the HTML URL with the header Accept: text/markdown' — CHANGED from partial"
- Content efficiency: partial — "homepage readable-text÷HTML = 14,125/464,197 bytes = 3.04% — a Next.js app-shell+RSC payload; above the 1% JS-shell floor, below the 5% server-rendered bar"
- Docs discoverable: pass — "'Docs' is a first-class header nav item on every page; docs root serves a 335-URL llms.html index and its own /docs/llms.txt (64KB)"
- Pricing machine readable: pass — "structured per-tier plans (Free/$0·$5 credit, Hobby/$10, Pro/$50, Enterprise/custom) + a per-second per-machine compute matrix + per-run fee, in clean markdown"
- Anti bot: pass — "No challenge on any public content: every fetch 200 (curl + the browser design pass); a cookie-consent banner is present but not a bot wall"

### Drive: 10.5 of 11 signals present

10.5/11 signals (10 full, 1 partial)

- Public API: pass — "Documented management REST API (trigger/manage runs, batches, bulk actions, errors, queues, schedules) via @trigger.dev/sdk + HTTP"
- Openapi spec: pass — "https://trigger.dev/docs/v3-openapi.yaml — fetchable OpenAPI 3.1.0, 248KB, verified 200 on 2026-10-15 (CHANGED from fail: the docs API surface links it; the apex /openapi.json still 404s, which is where the 2026-06 scan stopped)"
- MCP server: pass — "npm-installable MCP server with docs + interactive installer: 'npx trigger.dev@latest install-mcp' (v3 rubric full credit; /.well-known/mcp 404s but is not required)"
- SDK packages: pass — "Official npm packages: @trigger.dev/sdk (4.7.4) + trigger.dev CLI + React Realtime hooks (registry.npmjs.org verified)"
- Auth docs: pass — "/docs/apikeys: per-environment secret keys (TRIGGER_SECRET_KEY); public API keys omit payload/output for security"
- Agent self serve creds: partial — "Signup → dashboard/CLI key (no card required, but a human OAuth/signup step; programmatic key issuance not documented)"
- Webhooks: pass — "wait-for-token: 'pause for approvals, webhooks, and external events' — the waitToken URL is designed to be handed to external services (openapi spec) + docs/wait-for-token"
- CLI tool: pass — "Official trigger.dev CLI (init, dev, deploy, login, skills, install-mcp; command docs per page in docs index)"
- Error model: pass — "Documented JSON error model: OpenAPI components.schemas.ErrorResponse on 401/404/422/429 + a whole Errors API (error groups, resolve/ignore/unresolve) — CHANGED from partial (2026-06: schema unverified)"
- Rate limit docs: pass — "/docs/limits: 1,500 API requests/min all tiers (higher on request, paid), batchTrigger to 1,000 tasks/call, TTL/quota tables"
- Agent config: pass — "Multiple agent-config surfaces: trigger.dev/SKILL.md (bootstrap skill), /docs/skills (open-standard installable skills), /docs/mcp-agent-rules, MCP installer for 12 named clients"

## Coding harnesses

- Claude Code (Anthropic) — "Client Configuration … Claude Code: npx trigger.dev@latest install-mcp --client claude-code (or .mcp.json config 'command': 'npx', 'args': ['trigger.dev@latest', 'mcp'])" [source](https://trigger.dev/docs/mcp-introduction)
- Codex CLI (OpenAI) — "Codex CLI install tab: npx trigger.dev@latest install-mcp --client codex" [source](https://trigger.dev/docs/mcp-introduction)
- Cursor (Anysphere) — "Cursor install tab with .cursor/mcp.json config" [source](https://trigger.dev/docs/mcp-introduction)
- Windsurf (Codeium) — "Windsurf install tab: npx trigger.dev@latest install-mcp --client windsurf" [source](https://trigger.dev/docs/mcp-introduction)
- Cline (Cline) — "Cline install tab in the MCP client configuration list" [source](https://trigger.dev/docs/mcp-introduction)
- Zed Agent (Zed) — "Zed install tab in the MCP client configuration list" [source](https://trigger.dev/docs/mcp-introduction)
- Gemini CLI (Google) — "Gemini CLI install tab: npx trigger.dev@latest install-mcp --client gemini-cli" [source](https://trigger.dev/docs/mcp-introduction)
- Amp (Sourcegraph) — "AMP install tab in the MCP client configuration list" [source](https://trigger.dev/docs/mcp-introduction)
- OpenCode (opencode) — "opencode install tab in the MCP client configuration list" [source](https://trigger.dev/docs/mcp-introduction)
- Crush — "Crush install tab in the MCP client configuration list" [source](https://trigger.dev/docs/mcp-introduction)
- Ruler — "Ruler install tab in the MCP client configuration list" [source](https://trigger.dev/docs/mcp-introduction)

## Compliance & regions

- SOC 2 Type II: certified (scope: hosted) — "Trigger.dev is SOC 2 Type II and GDPR compliant." [source](https://trigger.dev/security)
- GDPR: self-attested (scope: hosted) — "SOC 2 Type II and GDPR compliant" [source](https://trigger.dev/security)
- HIPAA: enterprise-only (scope: hosted) — "Request a HIPAA Business Associate Agreement as a paid add-on to process Protected Health Information." [source](https://trigger.dev/security)
- Regions note: Region placement is now a real user-facing capability ('Deploy tasks to multiple regions for lower latency to users or data sources'; per-region static IPs) — so multi-region scores pass — but the actual region list and per-region addresses stay behind the dashboard (the auth-gated 'Regions' page), not the public surface. Sub-processors (updated 2026-10-13) show AWS + DigitalOcean (USA) and Cloudflare/ClickHouse as the infra plane; PostHog (EU) and some UK/US SaaS providers. Self-host placement is the operator's choice.

## Open source

- Repository: https://github.com/triggerdotdev/trigger.dev
- License: Apache-2.0
- Stars: 16545
- Forks: 1499
- Self-host methods: docker, compose, kubernetes, helm
- Self-host dependencies: postgres, redis, clickhouse, object-storage

## Company & presence

- Headquarters: Altrincham, Cheshire, United Kingdom (registered office: 3rd Floor, 1 Ashley Road, WA14 2DT); team UK remote/hybrid
- Founded: 2022
- Team size: small startup (~10–15; not formally disclosed)
- Funding: Seed, $3M, 2023-08
- Funding: Series A, $16M, 2025-12
- Privacy policy: https://trigger.dev/legal/privacy
- Terms of service: https://trigger.dev/legal
- GitHub: https://github.com/triggerdotdev/trigger.dev
- Blog: https://trigger.dev/blog
- Changelog: https://trigger.dev/changelog
- X/Twitter: https://twitter.com/triggerdotdev
- Discord: https://trigger.dev/discord
- LinkedIn: https://www.linkedin.com/company/triggerdotdev

## Nearby in the atlas

- [Claude Managed Agents](/products/claude-managed-agents.md)
- [n8n](/products/n8n.md)
- [Windmill](/products/windmill.md)
- [AgentNest](/products/agentnest.md)
- [Anthropic Sandbox Runtime (srt)](/products/sandbox-runtime.md)
- [Aptible](/products/aptible.md)

## Method

Facts come from the product's own public pages. Each claim carries its evidence: a quote and the page it came from. We count capabilities rather than grade products, and "Not assessed" means we have not checked, never that the product lacks it.