Storage

Query workflow runs, steps, hooks, and the underlying event log via the World storage interface.

The World storage interface exposes four sub-interfaces for querying workflow data:

world.events — The append-only event log. This is the source of truth for all workflow state. See Event Sourcing for background.
world.runs, world.steps, world.hooks — Materialized views derived from the event log, provided as convenience accessors for the most common query patterns.

import { getWorld } from "workflow/runtime";

const world = await getWorld();

The event log drives all workflow state. Use it for audit trails, debugging, and programmatic run cancellation.

events.create()

Create a new event for a workflow run. Most commonly used to cancel a run.

await world.events.create(runId, { 
  eventType: "run_cancelled", 
});

Parameter	Type	Description
`runId`	`string \| null`	The workflow run ID (`null` only for `run_created` events, where the server generates an ID)
`data`	`CreateEventRequest`	Event data including `eventType`
`params`	`object`	Optional parameters

Returns: EventResult — The created event and the affected entity (run/step/hook)

events.get()

Retrieve a single event by run ID and event ID.

const event = await world.events.get(runId, eventId);

Parameter	Type	Description
`runId`	`string`	The workflow run ID
`eventId`	`string`	The event ID

Returns: Event

events.list()

List events for a run with cursor pagination.

const result = await world.events.list({ runId, pagination: { cursor } });

Parameter	Type	Description
`params.runId`	`string`	Filter events by run ID
`params.pagination.cursor`	`string`	Cursor for the next page

Returns: { data: Event[], cursor?: string }

events.listByCorrelationId()

List events that share a correlation ID, useful for tracing related events across runs.

const result = await world.events.listByCorrelationId({ 
  correlationId: "order-123",
});

Parameter	Type	Description
`params.correlationId`	`string`	The correlation ID to filter by
`params.pagination.cursor`	`string`	Cursor for the next page

Returns: { data: Event[], cursor?: string }

Event Types

Category	Types
Run	`run_created`, `run_started`, `run_completed`, `run_failed`, `run_cancelled`
Step	`step_created`, `step_started`, `step_completed`, `step_failed`, `step_retrying`
Hook	`hook_created`, `hook_received`, `hook_disposed`, `hook_conflict`
Wait	`wait_created`, `wait_completed`

Materialized from run events. Use it to list and inspect workflow runs.

runs.get()

const run = await world.runs.get(runId);

Parameter	Type	Description
`runId`	`string`	The workflow run ID
`params.resolveData`	`'all' \| 'none'`	Whether to include input/output data. Default: `'all'`

Returns: WorkflowRun (or WorkflowRunWithoutData when resolveData: 'none')

runs.list()

const result = await world.runs.list({ 
  pagination: { cursor },
});

Parameter	Type	Description
`params.pagination.cursor`	`string`	Cursor for the next page
`params.resolveData`	`'all' \| 'none'`	Whether to include input/output data

Returns: { data: WorkflowRun[], cursor?: string }

Cancelling Runs

To cancel a run, create a run_cancelled event via world.events.create() (see world.events above), or use the CLI or Web UI helpers.

WorkflowRun Type

Field	Type	Description
`runId`	`string`	Unique run identifier
`status`	`string`	`'running'`, `'completed'`, `'failed'`, `'cancelled'`
`workflowName`	`string`	Machine-readable workflow identifier
`input`	`any`	Workflow input data (when `resolveData: 'all'`)
`output`	`any`	Workflow output data (when `resolveData: 'all'`)
`error`	`any`	Error data if the run failed
`startedAt`	`string`	ISO timestamp when the run started
`completedAt`	`string \| null`	ISO timestamp when the run completed

workflowName is a machine-readable identifier like workflow//./src/workflows/order//processOrder. Use parseWorkflowName() from workflow/observability to extract a display-friendly name.

Materialized from step events. Use it to list steps, inspect their input/output, and build progress dashboards.

steps.get()

const step = await world.steps.get(runId, stepId);

Parameter	Type	Description
`runId`	`string \| undefined`	The workflow run ID
`stepId`	`string`	The step ID
`params.resolveData`	`'all' \| 'none'`	Whether to include input/output data. Default: `'all'`

Returns: Step (or StepWithoutData when resolveData: 'none')

steps.list()

const result = await world.steps.list({ 
  runId,
  pagination: { cursor },
});

Parameter	Type	Description
`params.runId`	`string`	Filter steps by run ID
`params.pagination.cursor`	`string`	Cursor for the next page
`params.resolveData`	`'all' \| 'none'`	Whether to include input/output data

Returns: { data: Step[], cursor?: string }

Step Type

Field	Type	Description
`runId`	`string`	Parent workflow run ID
`stepId`	`string`	Unique step identifier
`stepName`	`string`	Machine-readable step identifier
`status`	`string`	`'running'`, `'completed'`, `'failed'`
`input`	`any`	Step input data (when `resolveData: 'all'`)
`output`	`any`	Step output data (when `resolveData: 'all'`)
`error`	`any`	Error data if the step failed
`attempt`	`number`	Current retry attempt number
`startedAt`	`string`	ISO timestamp when the step started
`completedAt`	`string \| null`	ISO timestamp when the step completed
`retryAfter`	`string \| null`	ISO timestamp for next retry attempt

Step I/O is serialized using the devalue format. Use hydrateResourceIO() from workflow/observability to deserialize it for display. See hydrateResourceIO().

stepName is a machine-readable identifier like step//./src/workflows/order//processPayment. Use parseStepName() from workflow/observability to extract the shortName for UI display.

Materialized from hook events. Hooks are pause points in workflows that wait for external input. Use this interface to look up hooks by ID or token, inspect metadata, and build UIs for pending approvals.

hooks.get()

const hook = await world.hooks.get(hookId);

Parameter	Type	Description
`hookId`	`string`	The hook ID

Returns: Hook

hooks.getByToken()

Look up a hook by its token. Useful in webhook resume flows where you receive a token in the callback URL.

const hook = await world.hooks.getByToken(token);

Parameter	Type	Description
`token`	`string`	The hook token

Returns: Hook

hooks.list()

const result = await world.hooks.list({ 
  pagination: { cursor },
});

Parameter	Type	Description
`params.pagination.cursor`	`string`	Cursor for the next page

Returns: { data: Hook[], cursor?: string }

Hook Type

Field	Type	Description
`runId`	`string`	Parent workflow run ID
`hookId`	`string`	Unique hook identifier
`token`	`string`	Hook token for resuming
`ownerId`	`string`	Owner (team/user) ID
`projectId`	`string`	Project ID
`environment`	`string`	Deployment environment
`metadata`	`object`	Custom metadata attached to the hook
`isWebhook`	`boolean`	Whether this is a webhook-style hook

import { getWorld } from "workflow/runtime";

const world = await getWorld();
let cursor: string | undefined;

const runs = await world.runs.list({ 
  pagination: { cursor },
}); 

cursor = runs.cursor; // pass to next call for pagination

Get a Run — Full Data vs. Metadata Only

import { getWorld } from "workflow/runtime";

const world = await getWorld();

// Full data (default) — includes serialized input/output
const run = await world.runs.get(runId); 

// Metadata only — lighter, no I/O loaded
const lightweight = await world.runs.get(runId, { 
  resolveData: "none", 
});

List Steps for a Progress Dashboard

import { getWorld } from "workflow/runtime";
import { parseStepName } from "workflow/observability"; 

const world = await getWorld();
const steps = await world.steps.list({ 
  runId,
  resolveData: "none",
}); 

const progress = steps.data.map((step) => {
  const parsed = parseStepName(step.stepName); 
  return {
    stepId: step.stepId,
    displayName: parsed?.shortName ?? step.stepName, 
    status: step.status,
  };
});

Hydrate Step I/O

import { getWorld } from "workflow/runtime";
import { hydrateResourceIO, observabilityRevivers } from "workflow/observability"; 

const world = await getWorld();
const step = await world.steps.get(runId, stepId); 
const hydrated = hydrateResourceIO(step, observabilityRevivers); 
console.log(hydrated.input, hydrated.output);

Cancel a Run

import { getWorld } from "workflow/runtime";

const world = await getWorld();
await world.events.create(runId, { 
  eventType: "run_cancelled", 
});

Look Up Hook by Token

import { getWorld } from "workflow/runtime";

const world = await getWorld();
const hook = await world.hooks.getByToken(token); 
console.log(hook.runId, hook.metadata);

List Events for Audit Trail

import { getWorld } from "workflow/runtime";

const world = await getWorld();
const events = await world.events.list({ runId }); 

for (const event of events.data) {
  console.log(event.eventType, event.createdAt);
}

Event Sourcing — How the event log powers workflow replay and state
getRun() — Higher-level API for working with individual runs
workflow/observability — Hydrate step I/O and parse display names
resumeHook() — Resume a workflow by sending a payload to a hook
Hooks — Core concepts for hooks and pause points
Workflows and Steps — Core concepts for steps

推荐订阅源

Workflow SDK Documentation