Inngest
Build and automate event-driven, serverless workflows effortlessly.. [Freemium]
Last verified:
What is Inngest?
Inngest is a developer-first, event-driven durable execution platform that enables building reliable, multi-step background jobs and serverless workflows without managing queues, infrastructure, or state. It replaces traditional queues, state management, and scheduling systems by allowing developers to write plain TypeScript, Python, or Go functions with built-in steps that automatically handle retries, delays, branching logic, and recovery from failures.
Key features include durable execution with automatic retries and exponential backoff, event-driven triggers from applications/Cron/webhooks/other functions, step-based workflows with up to 1000 steps per function, full observability with live traces and metrics, flow control with concurrency keys/throttling/batching/rate limiting, sleep durations up to a year (7 days on free plan), and the ability to run anywhere including edge, serverless, and traditional servers. The platform provides enterprise-grade reliability with SOC 2 compliance, end-to-end encryption, SSO/SAML support, and HIPAA BAA availability.
Inngest is designed for developers building AI agents and multi-step workflows, startups needing reliable background jobs, engineering teams migrating from complex orchestration systems, and enterprises requiring scalable workflow infrastructure. It's particularly valuable for teams building AI orchestration, email sequences, order processing, webhook retry systems, scheduled reports, and any multi-step code that needs to be unbreakable.
Inngest pricing
Pricing model: Freemium
Inngest offers three plans: Hobby ($0/mo) includes 50,000 executions, 5 concurrent steps, 50 realtime connections, 3 users, unlimited branch/staging envs, logs/traces/observability, basic alerting, and community support. Pro starts at $75/mo with 1M included executions (up to 20M add-on), 100+ concurrent steps, 1000+ realtime connections, 15+ users, granular metrics, increased scale/throughput, higher usage limits, and 7-day trace retention. Enterprise requires contacting sales and includes custom executions, 500-50k concurrent steps, 1000+ realtime connections, custom users, SAML/RBAC/audit trails, exportable observability, 90-day trace retention, and dedicated Slack channel. Usage-based pricing applies for executions beyond included amounts and events (first 1-5M events/day free, then $0.000050 per event for 1M-5M, decreasing with volume).
Inngest pros
- Automatic retries with exponential backoff on failure
- Zero infrastructure management required
- Durable execution persists state across failures
- Works with TypeScript, Python, and Go SDKs
- Runs anywhere - edge, serverless, traditional servers
- Up to 1000 steps per function
- Sleep durations up to a year (7 days free)
- Built-in observability with live traces and metrics
- Event-driven triggers from multiple sources
- Flow control with concurrency, throttling, batching
- SOC 2 compliant with regular security audits
- End-to-end encryption middleware available
- Replay failed workflows in bulk
- Bulk cancellation of thousands of runs
- Free hobby tier with 50,000 executions monthly
- Simple one-command CLI installation
- Local dev server with instant tracing
Inngest cons
- Step timeout maximum of 2 hours
- Step return payload limited to 4MB
- Function run state cannot exceed 32MB
- Free plan sleep limited to 7 days
- Event payload size 256KB on free tier (upgradable to 3MB)
- Maximum 5000 events per request
- Rate limiting period maximum is 24 hours
- Batch size hard limit of 10 MiB
- Rate limiting skips events instead of queueing them
- Requires HTTP invocation of functions via serve handler
- Learning curve for durable execution concepts
- Vendor lock-in to Inngest platform
Frequently asked questions about Inngest
What is Inngest and what problem does it solve?
Inngest is an event-driven durable execution platform that allows you to run fast, reliable code on any platform without managing queues, infrastructure, or state. It replaces queues, state management, and scheduling systems by enabling developers to write reliable multi-step code using plain TypeScript, Python, or Go functions. Inngest handles backend infrastructure, queueing, scaling, concurrency, throttling, rate limiting, and observability automatically.
How does durable execution work in Inngest?
Durable execution ensures workflows run until completion by using steps (step.run) that act as code-level transactions. Each step automatically retries on failure and only runs once on success. Function state is automatically managed for fault tolerance across steps. When a step fails, Inngest picks up where it left off, storing full event history for logging and future retries/replays without needing dead-letter queues.
What triggers can start an Inngest function?
Inngest functions are triggered asynchronously by events from multiple sources: your application backend using the Inngest SDK, Cron schedules using cron syntax, webhook events from any service, and another Inngest function for direct invocation. You can customize triggers with filtering for subset matching events, delaying execution to specific timestamps, batching multiple events, and using multiple trigger types in a single function.
What programming languages does Inngest support?
Inngest supports TypeScript (JavaScript), Python, and Go through official SDKs. You write functions in your preferred language and deploy to your favorite cloud provider. The SDKs provide simple building blocks and integrations with favorite tools, allowing you to define functions within your existing codebase.
Where can I deploy Inngest functions?
Inngest is platform-agnostic and runs anywhere including edge, serverless, and traditional servers. You can deploy to Vercel, Netlify, Cloudflare Pages, AWS Lambda, and other cloud providers in one click. Functions are invoked over HTTP by the Inngest platform via the SDK's serve handler, so you don't need to keep them always running.
What observability features does Inngest provide?
Inngest provides full observability without building it yourself, including structured logs, real-time traces with every prompt/response pair, live traces and metrics for AI workflows and agents, function and event logs, and granular metrics. The Hobby plan has 24-hour trace retention, Pro has 7-day retention, and Enterprise has 90-day retention. You can query, cancel, or replay runs without grepping logs across systems.
How does error handling work in Inngest?
Inngest automatically retries on error with exponential backoff. Each step is retried automatically on failure with configurable retry behavior. You can catch and throw specific errors to control retry behavior, including throwing InngestError with retry set to false to prevent retries. If something breaks, Inngest picks up where it left off, and you can recover from bugs by re-running failed workflows in bulk through replay functionality.
What are the usage limits for Inngest functions?
Key limits include: maximum 1000 steps per function, timeout up to 2 hours per step, step return payload limit of 4MB, function run state cannot exceed 32MB, event payload maximum of 256KB on free tier (upgradable to 3MB), maximum 5000 events per request, batch size hard limit of 10 MiB, and maximum rate limit period of 24 hours. Sleep duration goes up to a year (7 days on free plan).
How does flow control work in Inngest?
Flow control adds multi-tenant aware prioritization, concurrency with keys, throttling, batching, and rate limiting to Inngest functions. Concurrency with keys ensures fair resource distribution and eliminates noisy-neighbor issues. Throttling manages concurrency over time by enqueuing functions in a backlog. Rate limiting is a hard limit that skips events beyond a specific limit, ideal for preventing abuse. Batching processes multiple events in a single function for efficiency.
Is Inngest suitable for enterprise use?
Yes, Inngest provides enterprise-grade reliability and scalability with SOC 2 compliance through regular security audits, end-to-end encryption middleware, SSO and SAML support, RBAC (role-based access control), audit trails tracking all account actions including API calls and webhooks, dedicated Slack channel support, HIPAA BAA availability for healthcare compliance, and SLAs. The Enterprise plan includes custom executions, 500-50k concurrent steps, exportable observability, and 90-day trace retention.