Introduction

Durablex runs durable workflows: ordinary functions whose every step is saved as it runs, so the whole thing survives crashes, restarts, and deploys without losing progress or repeating work.

You write normal code. Durablex handles the retries, the waiting, and the resuming.

What it's for

Anything you don't want stuck in a single request/response cycle, or that has to keep going even when things fail:

• Sending emails or notifications, with retries.
• Calling an AI model or a flaky third-party API.
• Processing an upload, then doing follow-up work.
• Multi-step jobs that pause, wait, and continue - sometimes over days.

Example

const orderCreated = defineWorkflow<OrderData>({
  name: "order.created",
  retry: { maxAttempts: 3 },
  handler: async (ctx) => {
    const charge = await ctx.step.run("charge", () => chargeCard(ctx.event.data));
    await ctx.step.sleep("settle", "1h");
    return await ctx.step.run("ship", () => createShipment(charge));
  },
});

Each step.run runs once and its result is remembered. If ship fails, only ship retries - charge is never charged twice. If the process restarts during the one-hour wait, the run picks up exactly where it left off.

How it runs

Durablex Cloud runs the engine for you as a managed, multi-tenant service. You write workflows with the TypeScript SDK and connect a runner; the platform handles persistence, retries, scheduling, and the live console. No queue, database, or state machine to operate.

Next steps