Temporal Durable Execution Platform

Temporal is an open-source durable execution platform that enables developers to build applications where multi-step workflows survive process crashes, network failures, and service restarts without losing progress or requiring manual recovery. The platform captures the state of every workflow step as it executes, persisting it to a durable store so that if a process crashes mid-execution, the workflow resumes exactly where it left off upon restart. Developers write business logic as Workflow functions using SDKs for Python, Go, TypeScript, Ruby, C#, Java, and PHP, defining the sequence of operations as straightforward imperative code without explicit state machines, retry loops, or compensation handlers. Side-effect operations are encapsulated as Activities, which are functions that can call external APIs, access databases, and interact with unreliable systems, with automatic retry policies, timeout enforcement, and idempotency guarantees managed by the platform. The Temporal Service runs as infrastructure that manages workflow state, task queues, timers, signals for external communication, and query endpoints for state inspection, available as a self-hosted open-source deployment or as managed Temporal Cloud with enterprise SLAs. Common use cases include payment processing with compensating transactions, order fulfillment pipelines spanning multiple services, AI agent orchestration where LLM calls and tool executions are made durable, long-running jobs that execute for days or weeks, and human-in-the-loop workflows that pause indefinitely waiting for user input. The platform handles over twenty-one thousand GitHub stars and is MIT licensed. Customers include OpenAI, Salesforce, Twilio, Cloudflare, GitLab, NVIDIA, DoorDash, HashiCorp, and Snap, with testimonials citing dramatic reductions in incident response time and infrastructure complexity.