What problem does it solve? Long-running background workflows break when serverless functions time out or crash mid-execution, losing progress and duplicating work. This Skill teaches how to build durable, checkpointed workflows with Inngest step methods so each step retries independently and survives process restarts. ## Core Features & Use Cases - Durable Delays and Scheduling: Use step.sleep and step.sleepUntil to pause workflows for hours or days without consuming compute, ideal for cart abandonment reminders or scheduled follow-ups. - Event and Signal Coordination: Use step.waitForEvent and step.waitForSignal to pause for human approvals, webhook callbacks, or async API completions with timeouts and expression-based matching. - Function Composition and Parallelism: Use step.invoke to call other functions and await typed results, step.sendEvent for fan-out, and Promise.all for parallel step execution. - Use Case: Build a multi-step onboarding flow that sends a welcome email, waits three days, waits for the user to complete a profile event with a 7-day timeout, then invokes a billing setup function. ## Quick Start Ask the AI to write an Inngest function that sends a welcome email, sleeps for 24 hours, then waits for a user approval event with a timeout using step methods.
