BullMQ vs Inngest vs Trigger.dev for Node.js background jobs
BullMQ, Inngest, and Trigger.dev solve different operating problems. BullMQ gives a Node.js team a Redis-backed queue it runs itself. Inngest coordinates durable event-driven functions. Trigger.dev runs long-lived tasks through managed or self-hosted orchestration. The useful comparison is not a universal winner; it is which control plane, runtime, and billing unit fit the workload you actually have.
TL;DR verdict
Choose BullMQ when Redis ownership and queue-level control are acceptable. Choose Inngest when event-driven durable steps and managed orchestration fit the application. Choose Trigger.dev when long-running tasks, schedules, realtime status, and a cloud-or-self-hosted deployment model fit better. Prototype one representative retry-heavy workflow before committing.
Key takeaways
- BullMQ is a Redis-backed Node.js queue; rate limiting is application configuration rather than a hosted-plan quota.
- Inngest and Trigger.dev meter different units. Do not compare executions, steps, runs, requests, and concurrency as if they were interchangeable.
- Repository counters and npm distributions are dated adoption context, not evidence of unique users, reliability, or product quality.
- No controlled cross-product benchmark was run for this guide, so it does not rank throughput or cost efficiency.
At-a-glance comparison matrix
| Option | Product shape | Best fit | Commercial or operating model |
|---|---|---|---|
| BullMQ | Redis-backed distributed queues | Teams that want queue control and can operate Redis | Self-operated library |
| Inngest | Durable event-driven functions | Teams that want managed steps, waits, throttling, and traces | Hosted plans; self-hosting requires separate evaluation |
| Trigger.dev | Long-running background tasks | Teams that want managed or self-hosted task orchestration | Cloud plans plus compute/run charges |
Evidence-backed comparison
Pricing and operating model
BullMQ is an MIT-licensed self-hosted library, not a comparable hosted plan. Inngest currently lists Hobby at $0/month with 50,000 executions and Pro starting at $99/month with 1,000,000 executions. Trigger.dev currently lists Free at $0/month with $5 credits, Hobby at $10/month with $10 credits, and Pro at $50/month with $50 credits; managed compute and run charges are separate. These terms come from the Inngest and Trigger.dev pricing pages accessed 2026-08-25; check the current pages before selecting a plan.
Plans, quotas, and limits
Inngest Hobby currently includes 5 concurrent steps, while Trigger.dev documents plan concurrency separately and a limit of 1,500 API requests per minute across plans. BullMQ rate limiting is application configuration backed by Redis, not a vendor plan quota. Keep executions, steps, runs, requests, and concurrency as distinct units.
Adoption signals
npm's moving last-week endpoint reported 8,237,365 BullMQ downloads for 2026-08-19 through 2026-08-25 when rechecked 2026-08-26. The prior 2026-08-18 through 2026-08-24 snapshot reported Inngest 1,725,348 and @trigger.dev/sdk 942,453 downloads. On 2026-08-25, GitHub reported BullMQ 9,325 stars/671 forks, inngest-js 998/146, and trigger.dev 16,114/1,421. These dated snapshots use different windows and are not a current like-for-like ranking. Package distributions and repository counters are not unique users, growth, maturity, or migration evidence.
Versions and releases
On 2026-08-26, npm listed bullmq 6.3.0, and the matching package-specific GitHub release record identifies v6.3.0, published 2026-08-26. The package-specific v6.2.2 record remains available as the prior Node.js package release. The repository's unqualified latest-release endpoint instead returned vex2.2.0 from the Elixir package in the multi-language repository, so it is not the Node.js package version. The 2026-08-25 records for inngest and @trigger.dev/sdk were 4.18.1 and 4.5.12, with GitHub tags inngest@4.18.1 and v4.5.12. Recheck npm and package-specific release records before deployment; do not treat a monorepo-wide latest tag as the Node.js package version.
License
Current package records identify bullmq as MIT, inngest as Apache-2.0, and @trigger.dev/sdk as MIT. Trigger.dev's broader repository reports Apache-2.0, so distinguish the SDK package from the repository rather than calling every Trigger.dev artifact Apache-2.0. Do not characterize Inngest's complete hosted product or all repositories from the npm package license alone.
Runtime and integration fit
BullMQ documents a Redis-backed Node.js queue and its current npm package requires Node.js 14.17 or newer. Inngest's npm package requires Node.js 20 or newer and its docs identify TypeScript, Python, and Go workflows. @trigger.dev/sdk requires Node.js 18.20 or newer; Trigger.dev describes the SDK as framework-agnostic and publishes JavaScript framework guides. Keep compatibility claims to named versions and documented adapters.
Documented capabilities
BullMQ documents Redis-backed distributed queues, priorities, delayed/repeatable jobs, retries, concurrency, crash recovery, and parent-child dependencies. Inngest documents durable event-driven functions with queueing, scaling, concurrency, throttling, rate limiting, and observability. Trigger.dev documents long-running background tasks with queuing, retries, schedules, observability, and realtime status, in cloud or self-hosted deployments.
How to use popularity signals
Use dated npm distributions and repository counters only as adoption context. They do not establish category leadership, product quality, or migration causation. Keep recommendations conditional on Redis ownership, managed orchestration needs, deployment constraints, and licensing scope.
Source availability and current status
All cited npm, GitHub, documentation, and pricing endpoints were reachable on 2026-08-25. BullMQ's npm metadata, weekly download record, unqualified latest-release endpoint, v6.3.0 release record, and preserved v6.2.2 release record were rechecked on 2026-08-26. These retrievals establish source availability at access time only; they do not establish uptime, maturity, production readiness, continued maintenance, or managed-cloud reliability. Recheck source and service status at publication.
Decision framework
Start with control-plane ownership: BullMQ puts queue operation and Redis in your hands, while Inngest and Trigger.dev provide orchestration services with different self-hosting options. Then compare the job's maximum duration, retry and idempotency needs, side effects, schedule and realtime requirements, failure visibility, and data-retention rules. Estimate cost only after mapping the same workload to each product's execution, step, run, request, concurrency, and compute units. The final test should be one representative failure-prone job in the intended deployment environment.
Migration notes
Start with one job that has retries, a side effect, and an observable completion state. Record its idempotency key, timeout behavior, retry policy, failure alert, and data-retention needs. When moving from BullMQ, separate queue transport concerns from workflow state. When moving between managed orchestrators, map each provider's execution, step, run, request, and concurrency units before estimating cost.
Methodology
This comparison uses BullMQ, Inngest, and Trigger.dev documentation; npm package metadata and weekly download records; GitHub repository and release records; and the Inngest and Trigger.dev pricing pages. All were accessed 2026-08-25. BullMQ's npm, download, unqualified latest-release, current v6.3.0 release, and preserved v6.2.2 release records were rechecked 2026-08-26. Executions, steps, runs, requests, concurrency, and compute charges remain separate because the products meter different work. No controlled cross-product benchmark was run, so throughput and cost-efficiency claims would require exact versions, workload, environment, warmup, repetitions, and raw output. Recheck current prices, limits, versions, package-specific releases, and service status before choosing or deploying a product.
Source notes
- BullMQ documentation — accessed 2026-08-25
- BullMQ weekly npm downloads — accessed 2026-08-26
- BullMQ npm package metadata — accessed 2026-08-26
- BullMQ unqualified GitHub latest-release endpoint — accessed 2026-08-26
- BullMQ v6.2.2 GitHub release record — accessed 2026-08-26
- BullMQ v6.3.0 GitHub release record — accessed 2026-08-26
- BullMQ GitHub repository metadata — accessed 2026-08-25
- Inngest documentation — accessed 2026-08-25
- Inngest weekly npm downloads — accessed 2026-08-25
- Inngest npm package metadata — accessed 2026-08-25
- Inngest pricing — accessed 2026-08-25
- Inngest GitHub release record — accessed 2026-08-25
- Inngest GitHub repository metadata — accessed 2026-08-25
- Trigger.dev documentation — accessed 2026-08-25
- Trigger.dev SDK weekly npm downloads — accessed 2026-08-25
- Trigger.dev SDK npm package metadata — accessed 2026-08-25
- Trigger.dev pricing — accessed 2026-08-25
- Trigger.dev GitHub release record — accessed 2026-08-25
- Trigger.dev GitHub repository metadata — accessed 2026-08-25
Source-backed FAQ
Is BullMQ a hosted alternative to Inngest or Trigger.dev?
No. BullMQ is a Redis-backed library you operate. Inngest and Trigger.dev offer managed orchestration, so their plan rows and operational responsibilities are not directly equivalent.
Which option is fastest?
This guide does not declare one. A useful benchmark must pin exact versions, workload, runtime, hardware, Redis or deployment configuration, warmup, repetitions, and raw output.
What should a proof of concept test?
Test retries, idempotency, delayed work, concurrency controls, failure visibility, and the exact deployment environment you plan to use.