Temporal

Temporal is a durable execution platform for workflows that can't just crash and lose state: payment processing, order fulfillment, long-running agent tasks. Write your business logic as code, and Temporal guarantees it runs to completion even if servers crash, networks fail, or deployments happen mid-workflow.

What's free: The Temporal server is fully open source. MIT license. All core features are free: workflow execution, activity retries, timers, signals, queries, search, visibility.

Temporal's mental model is the big win: write normal-looking code (Go, Java, TypeScript, Python, .NET) and Temporal handles retries, timeouts, crash recovery, and state persistence automatically. No message queues to manage. No dead letter queues to monitor. Your workflow function just runs, no matter what.

The catch: operating the Temporal server is non-trivial. It needs Cassandra or PostgreSQL + Elasticsearch for visibility. A production cluster is 3+ nodes with careful resource planning. The programming model also requires understanding of deterministic constraints. Your workflow code can't use random numbers, current time, or non-deterministic operations, which trips up developers initially.