workflow（工作流）

*CRITICAL*: Always Use Correct workflow Documentation

Your knowledge of workflow is outdated.

The workflow documentation outlined below matches the installed version of the Workflow SDK. Follow these instructions before starting on any workflow-related tasks:

Search the bundled documentation in node_modules/workflow/docs/:

1. Find docs: glob "node_modules/workflow/docs/**/*.mdx"
2. Search content: grep "your query" node_modules/workflow/docs/

Documentation structure in node_modules/workflow/docs/:

• getting-started/ - Framework setup (next.mdx, express.mdx, hono.mdx, etc.)
• foundations/ - Core concepts (workflows-and-steps.mdx, hooks.mdx, streaming.mdx, etc.)
• api-reference/workflow/ - API docs (sleep.mdx, create-hook.mdx, fatal-error.mdx, etc.)
• api-reference/workflow-api/ - Client API (start.mdx, get-run.mdx, resume-hook.mdx, etc.)
• api-reference/workflow-api/world/ - World SDK (runs.mdx, steps.mdx, hooks.mdx, events.mdx, streams.mdx, queue.mdx, observability.mdx)
• ai/ - AI SDK integration docs
• errors/ - Error code documentation

Related packages also include bundled docs:

• @workflow/ai: node_modules/@workflow/ai/docs/ - DurableAgent and AI integration
• @workflow/core: node_modules/@workflow/core/docs/ - Core runtime (foundations, how-it-works)
• @workflow/next: node_modules/@workflow/next/docs/ - Next.js integration

When in doubt, update to the latest version of the Workflow SDK.

Official Resources

• Website: https://workflow-sdk.dev
• GitHub: https://github.com/vercel/workflow

Quick Reference

Directives:

"use workflow";  // First line - makes async function durable
"use step";      // First line - makes function a cached, retryable unit

Essential imports:

// Workflow primitives
import { sleep, fetch, createHook, createWebhook, getWritable } from "workflow";
import { FatalError, RetryableError } from "workflow";
import { getWorkflowMetadata, getStepMetadata } from "workflow";

// API operations
import { start, getRun, resumeHook, resumeWebhook } from "workflow/api";

// Observability & data hydration
import { hydrateResourceIO, observabilityRevivers, parseStepName, parseWorkflowName } from "workflow/observability";

// Framework integrations
import { withWorkflow } from "workflow/next";
import { workflow } from "workflow/vite";
import { workflow } from "workflow/astro";
// Or use modules: ["workflow/nitro"] for Nitro/Nuxt

// AI agent
import { DurableAgent } from "@workflow/ai/agent";

Prefer Step Functions to Avoid Sandbox Errors

"use workflow" functions run in a sandboxed VM. "use step" functions have full Node.js access. Put your logic in steps and use the workflow function purely for orchestration.

// Steps have full Node.js and npm access
async function fetchUserData(userId: string) {
  "use step";
  const response = await fetch(`https://api.example.com/users/${userId}`);
  return response.json();
}

async function processWithAI(data: any) {
  "use step";
  // AI SDK works in steps without workarounds
  return await generateText({
    model: openai("gpt-4"),
    prompt: `Process: ${JSON.stringify(data)}`,
  });
}

// Workflow orchestrates steps - no sandbox issues
export async function dataProcessingWorkflow(userId: string) {
  "use workflow";
  const data = await fetchUserData(userId);
  const processed = await processWithAI(data);
  return { success: true, processed };
}

Benefits: Steps have automatic retry, results are persisted for replay, and no sandbox restrictions.

Workflow Sandbox Limitations

When you need logic directly in a workflow function (not in a step), these restrictions apply:

Example - Using fetch in workflow context:

import { fetch } from "workflow";

export async function myWorkflow() {
  "use workflow";
  globalThis.fetch = fetch;  // Required for AI SDK and HTTP libraries
  // Now generateText() and other libraries work
}

Note: DurableAgent from @workflow/ai handles the fetch assignment automatically.

DurableAgent — AI Agents in Workflows

Use DurableAgent to build AI agents that maintain state and survive interruptions. It handles the workflow sandbox automatically (no manual globalThis.fetch needed).

import { DurableAgent } from "@workflow/ai/agent";
import { getWritable } from "workflow";
import { z } from "zod";
import type { UIMessageChunk } from "ai";

async function lookupData({ query }: { query: string }) {
  "use step";
  // Step functions have full Node.js access
  return `Results for "${query}"`;
}

export async function myAgentWorkflow(userMessage: string) {
  "use workflow";

  const agent = new DurableAgent({
    model: "anthropic/claude-sonnet-4-5",
    system: "You are a helpful assistant.",
    tools: {
      lookupData: {
        description: "Search for information",
        inputSchema: z.object({ query: z.string() }),
        execute: lookupData,
      },
    },
  });

  const result = await agent.stream({
    messages: [{ role: "user", content: userMessage }],
    writable: getWritable<UIMessageChunk>(),
    maxSteps: 10,
  });

  return result.messages;
}

Key points:

• getWritable<UIMessageChunk>() streams output to the workflow run's default stream
• Tool execute functions that need Node.js/npm access should use "use step"
• Tool execute functions that use workflow primitives (sleep(), createHook()) should NOT use "use step" — they run at the workflow level
• maxSteps limits the number of LLM calls (default is unlimited)
• Multi-turn: pass result.messages plus new user messages to subsequent agent.stream() calls

For more details on DurableAgent, check the AI docs in node_modules/@workflow/ai/docs/.

Starting Workflows & Child Workflows

Use start() to launch workflows from API routes. start() cannot be called directly in workflow context — wrap it in a step function.

import { start } from "workflow/api";

// From an API route — works directly
export async function POST() {
  const run = await start(myWorkflow, [arg1, arg2]);
  return Response.json({ runId: run.runId });
}

// No-args workflow
const run = await start(noArgWorkflow);

Starting child workflows from inside a workflow — must use a step:

import { start } from "workflow/api";

// Wrap start() in a step function
async function triggerChild(data: string) {
  "use step";
  const run = await start(childWorkflow, [data]);
  return run.runId;
}

export async function parentWorkflow() {
  "use workflow";
  const childRunId = await triggerChild("some data");  // Fire-and-forget via step
  await sleep("1h");
}

start() returns immediately — it doesn't wait for the workflow to complete. Use run.returnValue to await completion.

Hooks — Pause & Resume with External Events

Hooks let workflows wait for external data. Use createHook() inside a workflow and resumeHook() from API routes. Deterministic tokens are for createHook() + resumeHook() (server-side) only. createWebhook() always generates random tokens — do not pass a token option to createWebhook().

Single event

import { createHook } from "workflow";

export async function approvalWorkflow() {
  "use workflow";

  const hook = createHook<{ approved: boolean }>({
    token: "approval-123",  // deterministic token for external systems
  });

  const result = await hook;  // Workflow suspends here
  return result.approved;
}

Multiple events (iterable hooks)

Hooks implement AsyncIterable — use for await...of to receive multiple events:

import { createHook } from "workflow";

export async function chatWorkflow(channelId: string) {
  "use workflow";

  const hook = createHook<{ text: string; done?: boolean }>({
    token: `chat-${channelId}`,
  });

  for await (const event of hook) {
    await processMessage(event.text);
    if (event.done) break;
  }
}

Each resumeHook(token, payload) call delivers the next value to the loop.

Resuming from API routes

import { resumeHook } from "workflow/api";

export async function POST(req: Request) {
  const { token, data } = await req.json();
  await resumeHook(token, data);
  return new Response("ok");
}

Error Handling

Use FatalError for permanent failures (no retry), RetryableError for transient failures:

import { FatalError, RetryableError } from "workflow";

if (res.status >= 400 && res.status < 500) {
  throw new FatalError(`Client error: ${res.status}`);
}
if (res.status === 429) {
  throw new RetryableError("Rate limited", { retryAfter: "5m" });
}

Serialization

All data passed to/from workflows and steps must be serializable.

Supported built-in types: string, number, boolean, null, undefined, bigint, plain objects, arrays, Date, RegExp, URL, URLSearchParams, Map, Set, Headers, ArrayBuffer, typed arrays, Request, Response, ReadableStream, WritableStream.

Not supported: Functions, Symbols, WeakMap/WeakSet. Pass data, not callbacks.

Custom Class Serialization

Class instances can be serialized across workflow/step boundaries by implementing the @workflow/serde protocol. This is essential when a class has instance methods with "use step" or when you want to pass class instances between steps.

Install: @workflow/serde must be a dependency of the package containing the class.

Pattern: Add two static methods inside the class body using computed property syntax:

import { WORKFLOW_SERIALIZE, WORKFLOW_DESERIALIZE } from "@workflow/serde";

export class Point {
  x: number;
  y: number;

  constructor(x: number, y: number) {
    this.x = x;
    this.y = y;
  }

  // Serialize: return plain data (must be devalue-compatible types only)
  static [WORKFLOW_SERIALIZE](instance: Point) {
    return { x: instance.x, y: instance.y };
  }

  // Deserialize: reconstruct from plain data
  static [WORKFLOW_DESERIALIZE](data: { x: number; y: number }) {
    return new Point(data.x, data.y);
  }

  async computeDistance(other: Point) {
    "use step";
    return Math.sqrt((this.x - other.x) ** 2 + (this.y - other.y) ** 2);
  }
}

Critical rules:

1. Define serde methods INSIDE the class body as static methods with computed property syntax (static [WORKFLOW_SERIALIZE](...)). The SWC plugin detects them by scanning the class. Do NOT assign them externally (e.g., (MyClass as any)[WORKFLOW_SERIALIZE] =...) -- the compiler will not detect this.
2. Serde methods must return only devalue-compatible types (plain objects, arrays, primitives, Date, Map, Set, Uint8Array, etc.). No functions, no class instances, no Node.js-specific objects.
3. Add "use step" to Node.js-dependent instance methods. The SWC plugin strips "use step" method bodies from the workflow bundle. This is how you keep Node.js imports (fs, crypto, child_process, etc.) out of the workflow sandbox. The class shell with its serde methods remains in the workflow bundle; only the step method bodies are removed.
4. Do NOT manually register classes. The SWC plugin automatically generates registration code (an IIFE that sets classId and adds the class to the global registry). Manual calls to registerSerializationClass() are unnecessary and error-prone.
5. Do NOT use dynamic imports to work around sandbox restrictions. If a class method needs Node.js APIs, the correct solution is "use step", not /* @vite-ignore */ import(...).

When serde works well: Pure data classes, domain models, configuration objects, and classes where Node.js-dependent methods can be marked with "use step".

When to avoid serde: If a class is fundamentally inseparable from Node.js APIs (every method needs fs, net, etc.) and cannot meaningfully exist as a shell in the workflow sandbox, keep it entirely in step functions and pass plain data objects across boundaries instead.

Validating Serde Compliance

Use these tools to verify classes are correctly set up:

• workflow transform <file> --check-serde -- Shows the SWC transform output for a file and checks if serde classes are compliant (no Node.js imports remaining in the workflow bundle).
• workflow validate -- Scans all workflow files and reports serde compliance issues. Use --json for machine-readable output.
• SWC Playground -- The web playground at workbench/swc-playground shows a Serde Analysis panel when serde patterns are detected.
• Build-time warnings -- The builder automatically warns when serde classes have Node.js built-in imports remaining in the workflow bundle.

Streaming

Use getWritable() to stream data from workflows. getWritable() can be called in both workflow and step contexts, but you cannot interact with the stream (call getWriter(), write(), close()) directly in a workflow function. The stream must be passed to step functions for actual I/O, or steps can call getWritable() themselves.

Get the stream in a workflow, pass it to a step:

import { getWritable } from "workflow";

export async function myWorkflow() {
  "use workflow";
  const writable = getWritable();
  await writeData(writable, "hello world");
}

async function writeData(writable: WritableStream, chunk: string) {
  "use step";
  const writer = writable.getWriter();
  try {
    await writer.write(chunk);
  } finally {
    writer.releaseLock();
  }
}

Call getWritable() directly inside a step (no need to pass it):

import { getWritable } from "workflow";

async function streamData(chunk: string) {
  "use step";
  const writer = getWritable().getWriter();
  try {
    await writer.write(chunk);
  } finally {
    writer.releaseLock();
  }
}

Namespaced Streams

Use getWritable({namespace: 'name'}) to create multiple independent streams for different types of data. This is useful for separating logs from primary output, different log levels, agent outputs, metrics, or any distinct data channels. Long-running workflows benefit from namespaced streams because you can replay only the important events (e.g., final results) while keeping verbose logs in a separate stream.

Example: Log levels and agent output separation:

import { getWritable } from "workflow";

type LogEntry = { level: "debug" | "info" | "warn" | "error"; message: string; timestamp: number };
type AgentOutput = { type: "thought" | "action" | "result"; content: string };

async function logDebug(message: string) {
  "use step";
  const writer = getWritable<LogEntry>({ namespace: "logs:debug" }).getWriter();
  try {
    await writer.write({ level: "debug", message, timestamp: Date.now() });
  } finally {
    writer.releaseLock();
  }
}

async function logInfo(message: string) {
  "use step";
  const writer = getWritable<LogEntry>({ namespace: "logs:info" }).getWriter();
  try {
    await writer.write({ level: "info", message, timestamp: Date.now() });
  } finally {
    writer.releaseLock();
  }
}

async function emitAgentThought(thought: string) {
  "use step";
  const writer = getWritable<AgentOutput>({ namespace: "agent:thoughts" }).getWriter();
  try {
    await writer.write({ type: "thought", content: thought });
  } finally {
    writer.releaseLock();
  }
}

async function emitAgentResult(result: string) {
  "use step";
  // Important results go to the default stream for easy replay
  const writer = getWritable<AgentOutput>().getWriter();
  try {
    await writer.write({ type: "result", content: result });
  } finally {
    writer.releaseLock();
  }
}

export async function agentWorkflow(task: string) {
  "use workflow";

  await logInfo(`Starting task: ${task}`);
  await logDebug("Initializing agent context");
  await emitAgentThought("Analyzing the task requirements...");

  // ... agent processing ...

  await emitAgentResult("Task completed successfully");
  await logInfo("Workflow finished");
}

Consuming namespaced streams:

import { start, getRun } from "workflow/api";
import { agentWorkflow } from "./workflows/agent";

export async function POST(request: Request) {
  const run = await start(agentWorkflow, ["process data"]);

  // Access specific streams by namespace
  const results = run.getReadable({ namespace: undefined }); // Default stream (important results)
  const infoLogs = run.getReadable({ namespace: "logs:info" });
  const debugLogs = run.getReadable({ namespace: "logs:debug" });
  const thoughts = run.getReadable({ namespace: "agent:thoughts" });

  // Return only important results for most clients
  return new Response(results, { headers: { "Content-Type": "application/json" } });
}

// Resume from a specific point (useful for long sessions)
export async function GET(request: Request) {
  const { searchParams } = new URL(request.url);
  const runId = searchParams.get("runId")!;
  const startIndex = parseInt(searchParams.get("startIndex") || "0", 10);

  const run = getRun(runId);
  // Resume only the important stream, skip verbose debug logs
  const stream = run.getReadable({ startIndex });

  return new Response(stream);
}

Pro tip: For very long-running sessions (50+ minutes), namespaced streams help manage replay performance. Put verbose/debug output in separate namespaces so you can replay just the important events quickly.

Debugging

# Check workflow endpoints are reachable
npx workflow health
npx workflow health --port 3001  # Non-default port

# Visual dashboard for runs
npx workflow web
npx workflow web <run_id>

# CLI inspection (use --json for machine-readable output, --help for full usage)
npx workflow inspect runs
npx workflow inspect run <run_id>

# For Vercel-deployed projects, specify backend and project
npx workflow inspect runs --backend vercel --project <project-name> --team <team-slug>
npx workflow inspect run <run_id> --backend vercel --project <project-name> --team <team-slug>

# Open Vercel dashboard in browser for a specific run
npx workflow inspect run <run_id> --web
npx workflow web <run_id> --backend vercel --project <project-name> --team <team-slug>

# Cancel a running workflow
npx workflow cancel <run_id>
npx workflow cancel <run_id> --backend vercel --project <project-name> --team <team-slug>
# --env defaults to "production"; use --env preview for preview deployments

Debugging tips:

• Use --json (-j) on any command for machine-readable output
• Use --web to open the Vercel Observability dashboard in your browser
• Use --help on any command for full usage details
• Only import workflow APIs you actually use. Unused imports can cause 500 errors.

Testing Workflows

Workflow SDK provides a Vitest plugin for testing workflows in-process — no running server required.

Unit testing steps: Steps are just functions; without the compiler, "use step" is a no-op. Test them directly:

import { describe, it, expect } from "vitest";
import { createUser } from "./user-signup";

describe("createUser step", () => {
  it("should create a user", async () => {
    const user = await createUser("test@example.com");
    expect(user.email).toBe("test@example.com");
  });
});

Integration testing: Use @workflow/vitest for workflows using sleep(), hooks, webhooks, or retries:

// vitest.integration.config.ts
import { defineConfig } from "vitest/config";
import { workflow } from "@workflow/vitest";

export default defineConfig({
  plugins: [workflow()],
  test: {
    include: ["**/*.integration.test.ts"],
    testTimeout: 60_000,
  },
});

// approval.integration.test.ts
import { describe, it, expect } from "vitest";
import { start, getRun, resumeHook } from "workflow/api";
import { waitForHook, waitForSleep } from "@workflow/vitest";
import { approvalWorkflow } from "./approval";

describe("approvalWorkflow", () => {
  it("should publish when approved", async () => {
    const run = await start(approvalWorkflow, ["doc-123"]);

    // Wait for the hook, then resume it
    await waitForHook(run, { token: "approval:doc-123" });
    await resumeHook("approval:doc-123", { approved: true, reviewer: "alice" });

    // Wait for sleep, then wake it up
    const sleepId = await waitForSleep(run);
    await getRun(run.runId).wakeUp({ correlationIds: [sleepId] });

    const result = await run.returnValue;
    expect(result).toEqual({ status: "published", reviewer: "alice" });
  });
});

Testing webhooks: Use resumeWebhook() with a Request object — no HTTP server needed:

import { start, resumeWebhook } from "workflow/api";
import { waitForHook } from "@workflow/vitest";

const run = await start(ingestWorkflow, ["ep-1"]);
const hook = await waitForHook(run);  // Discovers the random webhook token
await resumeWebhook(hook.token, new Request("https://example.com/webhook", {
  method: "POST",
  body: JSON.stringify({ event: "order.created" }),
}));

Key APIs:

• start() — trigger a workflow
• run.returnValue — await workflow completion
• waitForHook(run, {token?}) / waitForSleep(run) — wait for workflow to reach a pause point
• resumeHook(token, data) / resumeWebhook(token, request) — resume paused workflows
• getRun(runId).wakeUp({correlationIds}) — skip sleep() calls

Best practices:

• Keep unit tests (no plugin) and integration tests (workflow() plugin) in separate configs
• Use deterministic hook tokens based on test data for easier resumption
• Set generous testTimeout — workflows may run longer than typical unit tests
• vi.mock() does not work in integration tests — step dependencies are bundled by esbuild

Observability & World SDK

Use await getWorld() to build observability dashboards, admin panels, and inspect workflow state. getWorld() is asynchronous and returns Promise<World> (dynamic import / env-based setup).

Key imports:

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

Key docs (grep node_modules/workflow/docs/ for full details):

• api-reference/workflow-api/world/storage.mdx — events, runs, steps, hooks (events are source of truth; others are materialized views)
• api-reference/workflow-api/world/observability.mdx — hydration, parsing, encryption

World SDK Method Signatures

⚠️ Pagination is nested: {pagination: {cursor}} — NOT {cursor} directly.

const world = await getWorld();

// Runs
const { data, cursor } = await world.runs.list({ pagination: { cursor }, resolveData: 'all' | 'none' });
const run = await world.runs.get(runId, { resolveData: 'all' | 'none' });
// Cancel via event creation (no cancel() method on runs)
await world.events.create(runId, { eventType: 'run_cancelled' });

// Steps — runId is top-level, NOT inside pagination
const { data, cursor } = await world.steps.list({ runId, pagination: { cursor }, resolveData: 'all' | 'none' });
const step = await world.steps.get(runId, stepId, { resolveData: 'all' | 'none' });

// Events
const { data, cursor } = await world.events.list({ runId, pagination: { cursor } });
await world.events.create(runId, { eventType: 'run_cancelled' });

// Hooks
const hook = await world.hooks.get(hookId);
const hook = await world.hooks.getByToken(token);

// Streams (methods on world.streams)
await world.streams.write(runId, name, chunk);
await world.streams.writeMulti?.(runId, name, chunks);
const readable = await world.streams.get(runId, name, startIndex);
await world.streams.close(runId, name);
const streamNames = await world.streams.list(runId);
const chunks = await world.streams.getChunks(runId, name, { limit, cursor });
const info = await world.streams.getInfo(runId, name);

// Queue (methods live directly on world — internal SDK infrastructure)
await world.queue(queueName, payload, opts);
const deploymentId = await world.getDeploymentId();

resolveData Parameter

Controls whether input/output data is included in the response. Accepts 'all' (default) or 'none'.

IMPORTANT: Even with 'all', data is still devalue-serialized. You MUST call hydrateResourceIO() to get usable JS values.

• Use 'none' for status polling, progress dashboards, run listings
• Use 'all' (or omit) when you need to inspect actual step I/O data — then always hydrate

// Lightweight status check — no I/O loaded
const run = await world.runs.get(runId, { resolveData: 'none' });
console.log(run.status); // 'running' | 'completed' | 'failed' | 'cancelled'

// Full inspection — resolveData includes data, hydrateResourceIO deserializes it
const step = await world.steps.get(runId, stepId); // defaults to 'all'
const hydrated = hydrateResourceIO(step, observabilityRevivers);

Common mistake: Checking step.input!== undefined after resolveData: 'all' and assuming the data is ready to use. The data exists but is serialized — always hydrate first.

Data Hydration (Devalue Format)

Step I/O is serialized via devalue with a 4-byte format prefix (devl). Without hydration, input/output are Uint8Array-like objects with numeric keys: {"0":100,"1":101,"2":118,"3":108,...} — these are NOT usable values.

Always hydrate before using I/O data:

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

const { data: steps } = await world.steps.list({ runId, resolveData: 'all' });
const hydrated = steps.map(s => hydrateResourceIO(s, observabilityRevivers));
// hydrated[0].input → [123, 2] (actual function arguments)
// hydrated[0].output → 125 (actual return value)

hydrateResourceIO works on both Step and WorkflowRun objects. For encrypted workflows, use getEncryptionKeyForRun() + hydrateResourceIOWithKey().

Name Parsing

parseWorkflowName(), parseStepName(), and parseClassName() return {shortName: string, moduleSpecifier: string} | null. Always use optional chaining:

const parsed = parseWorkflowName("workflow//./src/workflows/order//processOrder");
// parsed?.shortName → "processOrder"
// parsed?.moduleSpecifier → "./src/workflows/order"
// ⚠️ Returns null if format doesn't match

Event Types

Events are the append-only source of truth. Runs/Steps/Hooks are materialized views.

Error Handling Patterns

Three error strategies for different failure modes:

import { FatalError, RetryableError } from "workflow";

// Permanent failure — workflow terminates
throw new FatalError("Invalid input: missing required field");

// Transient failure — will retry
throw new RetryableError("API rate limited", { retryAfter: "5m" });

// Mixed criticality parallel execution
const results = await Promise.allSettled([
  criticalStep(data),    // Must succeed
  optionalStep(data),    // OK to fail
  enrichmentStep(data),  // OK to fail
]);
const [critical, optional, enrichment] = results;
if (critical.status === "rejected") throw new FatalError(critical.reason);