Add workflow versioning docs

[pranaygp]

Summary

Adds a new Foundations page explaining Workflow SDK versioning semantics for both v5 and v4 docs.

The page covers:

• default run pinning to the deployment that started the run
• canceling and rerunning affected in-flight runs on the latest deployment
• explicit self-upgrade patterns using deploymentId: "latest"
• carrying serializable context such as streams and abort signals across child runs

This also links existing versioning and deploymentId: "latest" mentions back to the new Foundations page, and calls out that deploymentId is currently a Vercelism that may become a more generic version concept in the World spec later.

Follow-up cookbook pass added targeted callouts for long-lived or multi-turn patterns:

• scheduled workflows
• DurableAgent
• AI SDK multi-turn sessions
• Chat SDK sessions
• Sandbox sessions

The docs now also distinguish v5 from v4 for child-run starts. Native start() from workflow context was added by e295bae41 / #1491, which is on main but not stable, so v5 examples encourage direct workflow-context start() and v4 keeps the step-wrapped pattern.

Validation

• git diff --check
• Parsed the updated v4/v5 Foundations meta.json files as JSON
• Verified v4/v5 copies of the touched cookbook pages match before the v5-only native-start pass
• Searched docs/content for stale v5 step-wrapped start() guidance and v4 direct-start guidance
• Confirmed origin/stable lacks the 'use step' directive in packages/core/src/runtime/start.ts, while main has it via e295bae41 / feat: allow start() to be called directly inside workflow functions #1491

Full docs lint/typecheck were not run because this checkout does not have node_modules installed, so biome, vitest, and changeset are unavailable locally.

[vercel]

The latest updates on your projects. Learn more about Vercel for GitHub.

Project	Deployment	Actions	Updated (UTC)
example-nextjs-workflow-turbopack	Ready	Preview, Comment	May 18, 2026 10:49pm
example-nextjs-workflow-webpack	Ready	Preview, Comment	May 18, 2026 10:49pm
example-workflow	Ready	Preview, Comment	May 18, 2026 10:49pm
workbench-astro-workflow	Ready	Preview, Comment	May 18, 2026 10:49pm
workbench-express-workflow	Ready	Preview, Comment	May 18, 2026 10:49pm
workbench-fastify-workflow	Ready	Preview, Comment	May 18, 2026 10:49pm
workbench-hono-workflow	Ready	Preview, Comment	May 18, 2026 10:49pm
workbench-nitro-workflow	Ready	Preview, Comment	May 18, 2026 10:49pm
workbench-nuxt-workflow	Ready	Preview, Comment	May 18, 2026 10:49pm
workbench-sveltekit-workflow	Ready	Preview, Comment	May 18, 2026 10:49pm
workbench-tanstack-start-workflow	Ready	Preview, Comment	May 18, 2026 10:49pm
workbench-vite-workflow	Ready	Preview, Comment	May 18, 2026 10:49pm
workflow-docs	Ready	Preview, ✅ 11 resolved, Open in v0	May 18, 2026 10:49pm
workflow-swc-playground	Ready	Preview, Comment	May 18, 2026 10:49pm
workflow-tarballs	Ready	Preview, Comment	May 18, 2026 10:49pm
workflow-web	Ready	Preview, Comment	May 18, 2026 10:49pm

[changeset-bot]

🦋 Changeset detected

Latest commit: 7eff96c

The changes in this PR will be included in the next version bump.

This PR includes changesets to release 0 packages

When changesets are added to this PR, you'll see the packages that this PR includes changesets for and the associated semver types

Not sure what this means? Click here to learn what changesets are.

Click here if you're a maintainer who wants to add another changeset to this PR

[github-actions]

📊 Benchmark Results

📈 Comparing against baseline from main branch. Green 🟢 = faster, Red 🔺 = slower.

workflow with no steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
💻 Local	🥇 Nitro	0.031s (-28.3% 🟢)	1.005s (~)	0.974s	10	1.00x
💻 Local	Express	0.036s (-19.6% 🟢)	1.005s (~)	0.970s	10	1.15x
🐘 Postgres	Express	0.045s (-21.9% 🟢)	1.011s (~)	0.966s	10	1.47x
🐘 Postgres	Nitro	0.046s (-51.3% 🟢)	1.013s (-2.9%)	0.967s	10	1.50x

workflow with 1 step

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
💻 Local	🥇 Express	1.070s (-4.9%)	2.006s (~)	0.936s	10	1.00x
💻 Local	Nitro	1.071s (-5.3% 🟢)	2.006s (~)	0.935s	10	1.00x
🐘 Postgres	Express	1.079s (-5.9% 🟢)	2.009s (~)	0.930s	10	1.01x
🐘 Postgres	Nitro	1.081s (-5.2% 🟢)	2.010s (~)	0.929s	10	1.01x

workflow with 10 sequential steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Nitro	10.396s (-4.4%)	11.018s (~)	0.622s	3	1.00x
🐘 Postgres	Express	10.402s (-5.1% 🟢)	11.016s (~)	0.614s	3	1.00x
💻 Local	Express	10.411s (-4.7%)	11.023s (~)	0.612s	3	1.00x
💻 Local	Nitro	10.412s (-4.9%)	11.022s (~)	0.610s	3	1.00x

workflow with 25 sequential steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
💻 Local	🥇 Express	13.432s (-10.3% 🟢)	14.027s (-6.7% 🟢)	0.595s	5	1.00x
🐘 Postgres	Express	13.435s (-7.9% 🟢)	14.017s (-6.7% 🟢)	0.581s	5	1.00x
🐘 Postgres	Nitro	13.484s (-7.6% 🟢)	14.017s (-6.7% 🟢)	0.533s	5	1.00x
💻 Local	Nitro	13.489s (-10.4% 🟢)	14.027s (-12.5% 🟢)	0.538s	5	1.00x

workflow with 50 sequential steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
💻 Local	🥇 Nitro	11.875s (-29.2% 🟢)	12.022s (-29.4% 🟢)	0.147s	8	1.00x
💻 Local	Express	11.890s (-28.4% 🟢)	12.022s (-29.4% 🟢)	0.132s	8	1.00x
🐘 Postgres	Express	11.941s (-14.7% 🟢)	12.391s (-15.1% 🟢)	0.450s	8	1.01x
🐘 Postgres	Nitro	11.944s (-14.5% 🟢)	12.266s (-14.3% 🟢)	0.321s	8	1.01x

Promise.all with 10 concurrent steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Express	1.136s (-9.8% 🟢)	2.007s (~)	0.870s	15	1.00x
🐘 Postgres	Nitro	1.137s (-10.8% 🟢)	2.007s (~)	0.870s	15	1.00x
💻 Local	Nitro	1.172s (-28.2% 🟢)	2.006s (-3.3%)	0.834s	15	1.03x
💻 Local	Express	1.183s (-20.6% 🟢)	2.006s (~)	0.823s	15	1.04x

Promise.all with 25 concurrent steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Nitro	1.192s (-49.3% 🟢)	2.007s (-33.3% 🟢)	0.815s	15	1.00x
🐘 Postgres	Express	1.224s (-48.2% 🟢)	2.007s (-33.3% 🟢)	0.783s	15	1.03x
💻 Local	Express	1.703s (-42.3% 🟢)	2.005s (-41.9% 🟢)	0.303s	15	1.43x
💻 Local	Nitro	1.704s (-45.8% 🟢)	2.005s (-48.4% 🟢)	0.302s	15	1.43x

Promise.all with 50 concurrent steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Nitro	1.305s (-62.5% 🟢)	2.008s (-49.9% 🟢)	0.702s	15	1.00x
🐘 Postgres	Express	1.324s (-62.0% 🟢)	2.008s (-49.9% 🟢)	0.684s	15	1.01x
💻 Local	Express	4.638s (-44.4% 🟢)	5.179s (-42.6% 🟢)	0.541s	6	3.55x
💻 Local	Nitro	4.910s (-41.2% 🟢)	5.511s (-38.9% 🟢)	0.601s	6	3.76x

Promise.race with 10 concurrent steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Express	1.155s (-8.1% 🟢)	2.009s (~)	0.854s	15	1.00x
🐘 Postgres	Nitro	1.163s (-7.5% 🟢)	2.010s (~)	0.847s	15	1.01x
💻 Local	Express	1.386s (-26.8% 🟢)	2.006s (-15.2% 🟢)	0.620s	15	1.20x
💻 Local	Nitro	1.388s (-25.6% 🟢)	2.006s (-14.3% 🟢)	0.618s	15	1.20x

Promise.race with 25 concurrent steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Express	1.199s (-48.8% 🟢)	2.008s (-33.3% 🟢)	0.810s	15	1.00x
🐘 Postgres	Nitro	1.199s (-48.7% 🟢)	2.007s (-33.3% 🟢)	0.808s	15	1.00x
💻 Local	Express	1.828s (-41.6% 🟢)	2.222s (-40.9% 🟢)	0.394s	14	1.53x
💻 Local	Nitro	1.935s (-36.9% 🟢)	2.316s (-40.4% 🟢)	0.381s	13	1.61x

Promise.race with 50 concurrent steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Express	1.305s (-62.7% 🟢)	2.008s (-49.9% 🟢)	0.702s	15	1.00x
🐘 Postgres	Nitro	1.309s (-62.4% 🟢)	2.009s (-49.9% 🟢)	0.700s	15	1.00x
💻 Local	Express	5.446s (-38.1% 🟢)	5.849s (-36.9% 🟢)	0.403s	6	4.17x
💻 Local	Nitro	5.557s (-39.2% 🟢)	6.179s (-38.4% 🟢)	0.622s	6	4.26x

workflow with 10 sequential data payload steps (10KB)

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Express	0.418s (-50.1% 🟢)	1.006s (-1.6%)	0.588s	60	1.00x
🐘 Postgres	Nitro	0.459s (-44.0% 🟢)	1.006s (~)	0.547s	60	1.10x
💻 Local	Nitro	0.468s (-52.2% 🟢)	1.004s (-8.2% 🟢)	0.535s	60	1.12x
💻 Local	Express	0.469s (-52.4% 🟢)	1.004s (-6.7% 🟢)	0.535s	60	1.12x

workflow with 25 sequential data payload steps (10KB)

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Express	1.047s (-47.0% 🟢)	1.705s (-24.5% 🟢)	0.659s	53	1.00x
🐘 Postgres	Nitro	1.066s (-44.7% 🟢)	1.576s (-25.0% 🟢)	0.510s	58	1.02x
💻 Local	Nitro	1.185s (-61.0% 🟢)	2.006s (-46.6% 🟢)	0.821s	45	1.13x
💻 Local	Express	1.191s (-60.5% 🟢)	2.005s (-44.1% 🟢)	0.815s	45	1.14x

workflow with 50 sequential data payload steps (10KB)

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Express	2.030s (-49.1% 🟢)	2.427s (-44.4% 🟢)	0.397s	50	1.00x
🐘 Postgres	Nitro	2.128s (-48.2% 🟢)	2.660s (-42.2% 🟢)	0.532s	46	1.05x
💻 Local	Nitro	2.704s (-70.9% 🟢)	3.032s (-69.7% 🟢)	0.328s	40	1.33x
💻 Local	Express	2.719s (-70.5% 🟢)	3.057s (-69.5% 🟢)	0.338s	40	1.34x

workflow with 10 concurrent data payload steps (10KB)

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Express	0.163s (-42.4% 🟢)	1.005s (~)	0.843s	60	1.00x
🐘 Postgres	Nitro	0.178s (-37.0% 🟢)	1.006s (~)	0.827s	60	1.10x
💻 Local	Express	0.455s (-18.8% 🟢)	1.004s (~)	0.550s	60	2.80x
💻 Local	Nitro	0.469s (-22.4% 🟢)	1.004s (-1.7%)	0.535s	60	2.89x

workflow with 25 concurrent data payload steps (10KB)

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Express	0.286s (-43.9% 🟢)	1.006s (~)	0.720s	90	1.00x
🐘 Postgres	Nitro	0.295s (-40.5% 🟢)	1.006s (~)	0.711s	90	1.03x
💻 Local	Nitro	2.205s (-13.1% 🟢)	2.944s (-2.2%)	0.738s	31	7.71x
💻 Local	Express	2.226s (-11.4% 🟢)	2.943s (-2.2%)	0.717s	31	7.78x

workflow with 50 concurrent data payload steps (10KB)

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Express	0.582s (-28.9% 🟢)	1.007s (-1.0%)	0.425s	120	1.00x
🐘 Postgres	Nitro	0.634s (-19.7% 🟢)	1.006s (~)	0.372s	120	1.09x
💻 Local	Nitro	10.021s (-10.5% 🟢)	10.611s (-9.0% 🟢)	0.590s	12	17.21x
💻 Local	Express	10.042s (-10.3% 🟢)	10.694s (-10.4% 🟢)	0.652s	12	17.25x

Stream Benchmarks(includes TTFB metrics)

workflow with stream

💻 Local Development

World	Framework	Workflow Time	TTFB	Slurp	Wall Time	Overhead	Samples	vs Fastest
💻 Local	🥇 Express	1.132s (+468.5% 🔺)	2.005s (+99.6% 🔺)	0.011s (-13.2% 🟢)	2.018s (+98.2% 🔺)	0.886s	10	1.00x
💻 Local	Nitro	1.133s (+430.4% 🔺)	2.005s (+99.6% 🔺)	0.011s (-15.2% 🟢)	2.018s (+98.1% 🔺)	0.885s	10	1.00x
🐘 Postgres	Nitro	1.135s (+453.6% 🔺)	2.001s (+100.2% 🔺)	0.001s (-20.0% 🟢)	2.010s (+98.7% 🔺)	0.875s	10	1.00x
🐘 Postgres	Express	1.135s (+453.5% 🔺)	1.999s (+100.2% 🔺)	0.001s (-25.0% 🟢)	2.010s (+98.7% 🔺)	0.875s	10	1.00x

stream pipeline with 5 transform steps (1MB)

💻 Local Development

World	Framework	Workflow Time	TTFB	Slurp	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Express	1.492s (+136.8% 🔺)	2.002s (+98.9% 🔺)	0.003s (-30.4% 🟢)	2.023s (+97.8% 🔺)	0.531s	30	1.00x
🐘 Postgres	Nitro	1.519s (+143.4% 🔺)	2.004s (+99.0% 🔺)	0.004s (-2.5%)	2.027s (+98.2% 🔺)	0.507s	30	1.02x
💻 Local	Nitro	1.707s (+103.5% 🔺)	2.013s (+98.9% 🔺)	0.009s (-4.3%)	2.203s (+97.4% 🔺)	0.496s	28	1.14x
💻 Local	Express	1.923s (+154.0% 🔺)	2.011s (+95.4% 🔺)	0.009s (-8.2% 🟢)	2.422s (+132.9% 🔺)	0.499s	25	1.29x

10 parallel streams (1MB each)

💻 Local Development

World	Framework	Workflow Time	TTFB	Slurp	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Express	0.627s (-34.7% 🟢)	1.033s (-19.2% 🟢)	0.000s (-20.7% 🟢)	1.045s (-20.0% 🟢)	0.418s	58	1.00x
🐘 Postgres	Nitro	0.662s (-31.7% 🟢)	1.029s (-17.5% 🟢)	0.000s (-17.2% 🟢)	1.049s (-16.6% 🟢)	0.387s	58	1.05x
💻 Local	Express	1.365s (+11.5% 🔺)	2.015s (~)	0.000s (+40.0% 🔺)	2.018s (~)	0.652s	30	2.18x
💻 Local	Nitro	1.561s (+27.6% 🔺)	2.016s (~)	0.000s (+364.3% 🔺)	2.197s (+8.7% 🔺)	0.636s	28	2.49x

fan-out fan-in 10 streams (1MB each)

💻 Local Development

World	Framework	Workflow Time	TTFB	Slurp	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Express	1.228s (-30.7% 🟢)	1.869s (-14.2% 🟢)	0.000s (NaN%)	1.900s (-13.6% 🟢)	0.673s	32	1.00x
🐘 Postgres	Nitro	1.325s (-26.1% 🟢)	1.997s (-6.7% 🟢)	0.000s (+86.7% 🔺)	2.019s (-7.2% 🟢)	0.694s	30	1.08x
💻 Local	Express	3.158s (-8.9% 🟢)	4.028s (~)	0.001s (-33.3% 🟢)	4.030s (~)	0.873s	15	2.57x
💻 Local	Nitro	3.187s (-5.9% 🟢)	4.029s (~)	0.000s (-75.0% 🟢)	4.031s (~)	0.844s	15	2.60x

Summary

Fastest Framework by World

Winner determined by most benchmark wins

World	🥇 Fastest Framework	Wins
💻 Local	Express	12/21
🐘 Postgres	Express	17/21

Fastest World by Framework

Winner determined by most benchmark wins

Framework	🥇 Fastest World	Wins
Express	🐘 Postgres	16/21
Nitro	🐘 Postgres	17/21

Column Definitions

• Workflow Time: Runtime reported by workflow (completedAt - createdAt) - primary metric
• TTFB: Time to First Byte - time from workflow start until first stream byte received (stream benchmarks only)
• Slurp: Time from first byte to complete stream consumption (stream benchmarks only)
• Wall Time: Total testbench time (trigger workflow + poll for result)
• Overhead: Testbench overhead (Wall Time - Workflow Time)
• Samples: Number of benchmark iterations run
• vs Fastest: How much slower compared to the fastest configuration for this benchmark

Worlds:

• 💻 Local: In-memory filesystem world (local development)
• 🐘 Postgres: PostgreSQL database world (local development)
• ▲ Vercel: Vercel production/preview deployment
• 🌐 Turso: Community world (local development)
• 🌐 MongoDB: Community world (local development)
• 🌐 Redis: Community world (local development)
• 🌐 Jazz: Community world (local development)
• 🌐 Redis: Community world (local development)
• 🌐 Redis + BullMQ: Community world (local development)
• 🌐 Cloudflare: Community world (local development)
• 🌐 MySQL: Community world (local development)
• 🌐 Azure: Community world (local development)
• 🌐 NATS JetStream: Community world (local development)
• 🌐 Upstash: Community world (local development)

---

📋 View full workflow run

---

❌ Some benchmark jobs failed:

• Local: success
• Postgres: success
• Vercel: failure

Check the workflow run for details.

[github-actions]

🧪 E2E Test Results

❌ Some tests failed

Summary

	Passed	Failed	Skipped	Total
❌ ▲ Vercel Production	1196	4	219	1419
✅ 💻 Local Development	1587	0	219	1806
✅ 📦 Local Production	1458	0	219	1677
❌ 🐘 Local Postgres	1585	2	219	1806
✅ 🪟 Windows	129	0	0	129
✅ 📋 Other	727	0	176	903
Total	6682	6	1052	7740

❌ Failed Tests

▲ Vercel Production (4 failed)

astro (1 failed):

• AbortController abortFromStepWorkflow: step abort cancels an in-flight sibling step

hono (2 failed):

• outputStreamWorkflow no startIndex (reads all chunks)
• stepFunctionPassingWorkflow - step function references can be passed as arguments (without closure vars) | wrun_01KRYMPWTNKF0CE4BP27X7R946 | 🔍 observability

nextjs-webpack (1 failed):

• parallelSleepWorkflow | wrun_01KRYME5H0F7D8EH2AAF56T0Z2 | 🔍 observability

🐘 Local Postgres (2 failed)

nextjs-webpack-stable-lazy-discovery-disabled (1 failed):

• DurableAgent e2e core basic text response

nextjs-webpack-stable-lazy-discovery-enabled (1 failed):

• addTenWorkflow | wrun_01KRYMBT8QHFK9EZ9FDY03D9G0

Details by Category

❌ ▲ Vercel Production

App	Passed	Failed	Skipped
❌ astro	102	1	26
✅ example	103	0	26
✅ express	103	0	26
✅ fastify	103	0	26
❌ hono	101	2	26
✅ nextjs-turbopack	127	0	2
❌ nextjs-webpack	126	1	2
✅ nitro	103	0	26
✅ nuxt	103	0	26
✅ sveltekit	122	0	7
✅ vite	103	0	26

✅ 💻 Local Development

App	Passed	Failed	Skipped
✅ astro-stable	104	0	25
✅ express-stable	104	0	25
✅ fastify-stable	104	0	25
✅ hono-stable	104	0	25
✅ nextjs-turbopack-canary	110	0	19
✅ nextjs-turbopack-stable-lazy-discovery-disabled	129	0	0
✅ nextjs-turbopack-stable-lazy-discovery-enabled	129	0	0
✅ nextjs-webpack-canary	110	0	19
✅ nextjs-webpack-stable-lazy-discovery-disabled	129	0	0
✅ nextjs-webpack-stable-lazy-discovery-enabled	129	0	0
✅ nitro-stable	104	0	25
✅ nuxt-stable	104	0	25
✅ sveltekit-stable	123	0	6
✅ vite-stable	104	0	25

✅ 📦 Local Production

App	Passed	Failed	Skipped
✅ astro-stable	104	0	25
✅ express-stable	104	0	25
✅ fastify-stable	104	0	25
✅ hono-stable	104	0	25
✅ nextjs-turbopack-canary	110	0	19
✅ nextjs-turbopack-stable-lazy-discovery-disabled	129	0	0
✅ nextjs-webpack-canary	110	0	19
✅ nextjs-webpack-stable-lazy-discovery-disabled	129	0	0
✅ nextjs-webpack-stable-lazy-discovery-enabled	129	0	0
✅ nitro-stable	104	0	25
✅ nuxt-stable	104	0	25
✅ sveltekit-stable	123	0	6
✅ vite-stable	104	0	25

❌ 🐘 Local Postgres

App	Passed	Failed	Skipped
✅ astro-stable	104	0	25
✅ express-stable	104	0	25
✅ fastify-stable	104	0	25
✅ hono-stable	104	0	25
✅ nextjs-turbopack-canary	110	0	19
✅ nextjs-turbopack-stable-lazy-discovery-disabled	129	0	0
✅ nextjs-turbopack-stable-lazy-discovery-enabled	129	0	0
✅ nextjs-webpack-canary	110	0	19
❌ nextjs-webpack-stable-lazy-discovery-disabled	128	1	0
❌ nextjs-webpack-stable-lazy-discovery-enabled	128	1	0
✅ nitro-stable	104	0	25
✅ nuxt-stable	104	0	25
✅ sveltekit-stable	123	0	6
✅ vite-stable	104	0	25

✅ 🪟 Windows

App	Passed	Failed	Skipped
✅ nextjs-turbopack	129	0	0

✅ 📋 Other

App	Passed	Failed	Skipped
✅ e2e-local-dev-nest-stable	104	0	25
✅ e2e-local-dev-tanstack-start-	104	0	25
✅ e2e-local-postgres-nest-stable	104	0	25
✅ e2e-local-postgres-tanstack-start-	104	0	25
✅ e2e-local-prod-nest-stable	104	0	25
✅ e2e-local-prod-tanstack-start-	104	0	25
✅ e2e-vercel-prod-tanstack-start	103	0	26

---

📋 View full workflow run

---

❌ Some E2E test jobs failed:

• Vercel Prod: failure
• Local Dev: success
• Local Prod: failure
• Local Postgres: failure
• Windows: success

Check the workflow run for details.

[Copilot]

Pull request overview

Adds a new "Versioning" Foundations page (v4 and v5) explaining default run pinning to deployments, recovery via cancel + rerun, opting into newer code with deploymentId: "latest", and carrying serializable state/streams across child-run continuations. Existing references to deploymentId: "latest" and versioning behavior are linked back to the new page. Cookbook pages also pick up callouts about long-lived / multi-turn patterns and, on v5 only, are updated to reflect that start() can now be called directly inside "use workflow" functions (via #1491), while v4 retains the step-wrapped pattern.

Changes:

• New foundations/versioning.mdx for both v4 and v5, plus meta.json + index.mdx + linkbacks from vercel-world, start.mdx, and code-transform.mdx.
• v5 cookbook + migration guides updated to drop "use step" wrappers around start() in favor of calling start() directly from the workflow function; v4 keeps the step-wrapped pattern.
• Added "Versioning" callouts to scheduling, durable-agent, ai-sdk, chat-sdk, and sandbox cookbooks; fixed a few small wording/grammar issues in code-transform.mdx.

Reviewed changes

Copilot reviewed 32 out of 32 changed files in this pull request and generated no comments.

Show a summary per file

File	Description
docs/content/docs/v5/foundations/versioning.mdx	New Foundations page explaining run pinning, recovery, and explicit upgrade boundaries.
docs/content/docs/v4/foundations/versioning.mdx	v4 counterpart; uses step-wrapped continueDigest/continueExportOnLatest helpers.
docs/content/docs/v{4,5}/foundations/{meta.json,index.mdx}	Surface the new Versioning page in nav and the foundations index.
docs/content/docs/v{4,5}/deploying/world/vercel-world.mdx	Link the deployment-pinning section to the new Versioning page.
docs/content/docs/v{4,5}/how-it-works/code-transform.mdx	Link "atomic versioning" reference; fix "function assume" / "from being upgraded" wording.
docs/content/docs/v{4,5}/api-reference/workflow-api/start.mdx	Soften "Vercel-specific" to "Vercelism", mention possible version rename, link Versioning. v5 also notes native workflow-context start().
docs/content/docs/v5/foundations/starting-workflows.mdx	Note v5 can call start() from workflows; link Composition + Versioning.
docs/content/docs/v5/cookbook/common-patterns/workflow-composition.mdx	Replace step-wrapped child spawn with direct start() call.
docs/content/docs/v5/cookbook/advanced/child-workflows.mdx	Inline batched start() calls; rename spawnReportChunk → startReportChunk; update tips.
docs/content/docs/v{4,5}/cookbook/common-patterns/scheduling.mdx	Callout linking to Versioning for long-lived schedules.
docs/content/docs/v{4,5}/cookbook/agent-patterns/durable-agent.mdx	Callout linking to Versioning for multi-turn agents.
docs/content/docs/v{4,5}/cookbook/integrations/{ai-sdk,chat-sdk,sandbox}.mdx	Callouts linking to Versioning for long-lived sessions.
docs/content/docs/v4/cookbook/common-patterns/workflow-composition.mdx	Update deploymentId: "latest" callout wording + link to Versioning.
docs/content/docs/v4/cookbook/advanced/child-workflows.mdx	Link Versioning from the deploymentId: "latest" tip.
docs/content/docs/v4/changelog/eager-processing.mdx	Reword to "step-wrapped start() helper" to match v4 phrasing.
docs/content/docs/v5/migration-guides/{migrating-from-trigger-dev,migrating-from-temporal,migrating-from-inngest,migrating-from-aws-step-functions}.mdx	Update to v5 native workflow-context start(); only collect step retains "use step".

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[TooTallNate]

Comment-only review

The new versioning foundations page is a useful addition — the four sections (default pinning → recovery → self-upgrade via start({ deploymentId: "latest" }) → carrying serializable context forward) cover the model clearly, and the v4-vs-v5 differentiation (step-wrapped start() vs direct start() from workflow context, gated on whether PR #1491 has landed on the target branch) is correctly handled. The cross-links from cookbook patterns to the new foundations page tie everything together nicely.

Three observations — see inline.

Inline comments

1. Wording: this PR introduces "Vercelism" as a new term and uses it 6 times across the diff, including replacing existing "Vercel-specific feature" wording in two <Callout>s. "Vercelism" reads as informal slang and isn't part of the established docs vocabulary. The existing tone uses "Vercel-specific", "Vercel deployments", "Vercel World", "non-Vercel" — all precise. Suggested restoring the prior phrasing. This is the only thing I'd consider blocking.
2. Broken link in v4: versioning.mdx references /docs/foundations/cancellation in its related frontmatter, but that page is v5-only. Easy fix.
3. Missing changeset: convention for docs-only PRs is an empty ---\n--- file.

A few small things I noticed

• The pre-emptively check for a conflict with getHookByToken() example (line 96 of versioning.mdx) — nice that you noted the race window in a later doc (#2011 callout), but this page doesn't have that warning. Worth a cross-link or <Callout type="warn"> here too, since the recovery pattern uses both start() and deploymentId: "latest" together.
• The _idempotency / using syntax is great. Worth confirming the docs site's syntax highlighter renders it — I noticed #1988 recently fixed a homepage link lint regression, and the using syntax is TypeScript 5.2+ specific.
• dailyDigest example (line 137-152): the state parameter is DigestState but the workflow doesn't actually use state.userId after the first sendDigest call — pattern is fine, just noting if you wanted to make the example illustrate state evolution more vividly, having sendDigest read state.lastSentAt (e.g. to avoid sending duplicates) would make the carry-forward intent clearer.

Overall the page reads well — once the Vercelism wording is restored and the v4 link is fixed, this should be good to go.

[TooTallNate]

All three review comments addressed in 7eff96c — Vercelism wording reverted, v4 broken cancellation link fixed, empty changeset added. LGTM, approving.

[github-actions]

Backport PR opened against stable: #2014. Merge conflicts were resolved by AI — please review carefully.

[pranaygp]

Summary

Adds a new Foundations page explaining Workflow SDK versioning semantics for both v5 and v4 docs.

The page covers:

• default run pinning to the deployment that started the run
• canceling and rerunning affected in-flight runs on the latest deployment
• explicit self-upgrade patterns using deploymentId: "latest"
• carrying serializable context such as streams and abort signals across child runs

This also links existing versioning and deploymentId: "latest" mentions back to the new Foundations page, and calls out that deploymentId is currently a Vercelism that may become a more generic version concept in the World spec later.

Follow-up cookbook pass added targeted callouts for long-lived or multi-turn patterns:

• scheduled workflows
• DurableAgent
• AI SDK multi-turn sessions
• Chat SDK sessions
• Sandbox sessions

The docs now also distinguish v5 from v4 for child-run starts. Native start() from workflow context was added by e295bae41 / #1491, which is on main but not stable, so v5 examples encourage direct workflow-context start() and v4 keeps the step-wrapped pattern.

Validation

• git diff --check
• Parsed the updated v4/v5 Foundations meta.json files as JSON
• Verified v4/v5 copies of the touched cookbook pages match before the v5-only native-start pass
• Searched docs/content for stale v5 step-wrapped start() guidance and v4 direct-start guidance
• Confirmed origin/stable lacks the 'use step' directive in packages/core/src/runtime/start.ts, while main has it via e295bae41 / feat: allow start() to be called directly inside workflow functions #1491

Full docs lint/typecheck were not run because this checkout does not have node_modules installed, so biome, vitest, and changeset are unavailable locally.

[vercel]

The latest updates on your projects. Learn more about Vercel for GitHub.

Project	Deployment	Actions	Updated (UTC)
example-nextjs-workflow-turbopack	Ready	Preview, Comment	May 18, 2026 10:49pm
example-nextjs-workflow-webpack	Ready	Preview, Comment	May 18, 2026 10:49pm
example-workflow	Ready	Preview, Comment	May 18, 2026 10:49pm
workbench-astro-workflow	Ready	Preview, Comment	May 18, 2026 10:49pm
workbench-express-workflow	Ready	Preview, Comment	May 18, 2026 10:49pm
workbench-fastify-workflow	Ready	Preview, Comment	May 18, 2026 10:49pm
workbench-hono-workflow	Ready	Preview, Comment	May 18, 2026 10:49pm
workbench-nitro-workflow	Ready	Preview, Comment	May 18, 2026 10:49pm
workbench-nuxt-workflow	Ready	Preview, Comment	May 18, 2026 10:49pm
workbench-sveltekit-workflow	Ready	Preview, Comment	May 18, 2026 10:49pm
workbench-tanstack-start-workflow	Ready	Preview, Comment	May 18, 2026 10:49pm
workbench-vite-workflow	Ready	Preview, Comment	May 18, 2026 10:49pm
workflow-docs	Ready	Preview, ✅ 11 resolved, Open in v0	May 18, 2026 10:49pm
workflow-swc-playground	Ready	Preview, Comment	May 18, 2026 10:49pm
workflow-tarballs	Ready	Preview, Comment	May 18, 2026 10:49pm
workflow-web	Ready	Preview, Comment	May 18, 2026 10:49pm

[changeset-bot]

🦋 Changeset detected

Latest commit: 7eff96c

The changes in this PR will be included in the next version bump.

This PR includes changesets to release 0 packages

When changesets are added to this PR, you'll see the packages that this PR includes changesets for and the associated semver types

Not sure what this means? Click here to learn what changesets are.

Click here if you're a maintainer who wants to add another changeset to this PR

[github-actions]

📊 Benchmark Results

📈 Comparing against baseline from main branch. Green 🟢 = faster, Red 🔺 = slower.

workflow with no steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
💻 Local	🥇 Nitro	0.031s (-28.3% 🟢)	1.005s (~)	0.974s	10	1.00x
💻 Local	Express	0.036s (-19.6% 🟢)	1.005s (~)	0.970s	10	1.15x
🐘 Postgres	Express	0.045s (-21.9% 🟢)	1.011s (~)	0.966s	10	1.47x
🐘 Postgres	Nitro	0.046s (-51.3% 🟢)	1.013s (-2.9%)	0.967s	10	1.50x

workflow with 1 step

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
💻 Local	🥇 Express	1.070s (-4.9%)	2.006s (~)	0.936s	10	1.00x
💻 Local	Nitro	1.071s (-5.3% 🟢)	2.006s (~)	0.935s	10	1.00x
🐘 Postgres	Express	1.079s (-5.9% 🟢)	2.009s (~)	0.930s	10	1.01x
🐘 Postgres	Nitro	1.081s (-5.2% 🟢)	2.010s (~)	0.929s	10	1.01x

workflow with 10 sequential steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Nitro	10.396s (-4.4%)	11.018s (~)	0.622s	3	1.00x
🐘 Postgres	Express	10.402s (-5.1% 🟢)	11.016s (~)	0.614s	3	1.00x
💻 Local	Express	10.411s (-4.7%)	11.023s (~)	0.612s	3	1.00x
💻 Local	Nitro	10.412s (-4.9%)	11.022s (~)	0.610s	3	1.00x

workflow with 25 sequential steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
💻 Local	🥇 Express	13.432s (-10.3% 🟢)	14.027s (-6.7% 🟢)	0.595s	5	1.00x
🐘 Postgres	Express	13.435s (-7.9% 🟢)	14.017s (-6.7% 🟢)	0.581s	5	1.00x
🐘 Postgres	Nitro	13.484s (-7.6% 🟢)	14.017s (-6.7% 🟢)	0.533s	5	1.00x
💻 Local	Nitro	13.489s (-10.4% 🟢)	14.027s (-12.5% 🟢)	0.538s	5	1.00x

workflow with 50 sequential steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
💻 Local	🥇 Nitro	11.875s (-29.2% 🟢)	12.022s (-29.4% 🟢)	0.147s	8	1.00x
💻 Local	Express	11.890s (-28.4% 🟢)	12.022s (-29.4% 🟢)	0.132s	8	1.00x
🐘 Postgres	Express	11.941s (-14.7% 🟢)	12.391s (-15.1% 🟢)	0.450s	8	1.01x
🐘 Postgres	Nitro	11.944s (-14.5% 🟢)	12.266s (-14.3% 🟢)	0.321s	8	1.01x

Promise.all with 10 concurrent steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Express	1.136s (-9.8% 🟢)	2.007s (~)	0.870s	15	1.00x
🐘 Postgres	Nitro	1.137s (-10.8% 🟢)	2.007s (~)	0.870s	15	1.00x
💻 Local	Nitro	1.172s (-28.2% 🟢)	2.006s (-3.3%)	0.834s	15	1.03x
💻 Local	Express	1.183s (-20.6% 🟢)	2.006s (~)	0.823s	15	1.04x

Promise.all with 25 concurrent steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Nitro	1.192s (-49.3% 🟢)	2.007s (-33.3% 🟢)	0.815s	15	1.00x
🐘 Postgres	Express	1.224s (-48.2% 🟢)	2.007s (-33.3% 🟢)	0.783s	15	1.03x
💻 Local	Express	1.703s (-42.3% 🟢)	2.005s (-41.9% 🟢)	0.303s	15	1.43x
💻 Local	Nitro	1.704s (-45.8% 🟢)	2.005s (-48.4% 🟢)	0.302s	15	1.43x

Promise.all with 50 concurrent steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Nitro	1.305s (-62.5% 🟢)	2.008s (-49.9% 🟢)	0.702s	15	1.00x
🐘 Postgres	Express	1.324s (-62.0% 🟢)	2.008s (-49.9% 🟢)	0.684s	15	1.01x
💻 Local	Express	4.638s (-44.4% 🟢)	5.179s (-42.6% 🟢)	0.541s	6	3.55x
💻 Local	Nitro	4.910s (-41.2% 🟢)	5.511s (-38.9% 🟢)	0.601s	6	3.76x

Promise.race with 10 concurrent steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Express	1.155s (-8.1% 🟢)	2.009s (~)	0.854s	15	1.00x
🐘 Postgres	Nitro	1.163s (-7.5% 🟢)	2.010s (~)	0.847s	15	1.01x
💻 Local	Express	1.386s (-26.8% 🟢)	2.006s (-15.2% 🟢)	0.620s	15	1.20x
💻 Local	Nitro	1.388s (-25.6% 🟢)	2.006s (-14.3% 🟢)	0.618s	15	1.20x

Promise.race with 25 concurrent steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Express	1.199s (-48.8% 🟢)	2.008s (-33.3% 🟢)	0.810s	15	1.00x
🐘 Postgres	Nitro	1.199s (-48.7% 🟢)	2.007s (-33.3% 🟢)	0.808s	15	1.00x
💻 Local	Express	1.828s (-41.6% 🟢)	2.222s (-40.9% 🟢)	0.394s	14	1.53x
💻 Local	Nitro	1.935s (-36.9% 🟢)	2.316s (-40.4% 🟢)	0.381s	13	1.61x

Promise.race with 50 concurrent steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Express	1.305s (-62.7% 🟢)	2.008s (-49.9% 🟢)	0.702s	15	1.00x
🐘 Postgres	Nitro	1.309s (-62.4% 🟢)	2.009s (-49.9% 🟢)	0.700s	15	1.00x
💻 Local	Express	5.446s (-38.1% 🟢)	5.849s (-36.9% 🟢)	0.403s	6	4.17x
💻 Local	Nitro	5.557s (-39.2% 🟢)	6.179s (-38.4% 🟢)	0.622s	6	4.26x

workflow with 10 sequential data payload steps (10KB)

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Express	0.418s (-50.1% 🟢)	1.006s (-1.6%)	0.588s	60	1.00x
🐘 Postgres	Nitro	0.459s (-44.0% 🟢)	1.006s (~)	0.547s	60	1.10x
💻 Local	Nitro	0.468s (-52.2% 🟢)	1.004s (-8.2% 🟢)	0.535s	60	1.12x
💻 Local	Express	0.469s (-52.4% 🟢)	1.004s (-6.7% 🟢)	0.535s	60	1.12x

workflow with 25 sequential data payload steps (10KB)

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Express	1.047s (-47.0% 🟢)	1.705s (-24.5% 🟢)	0.659s	53	1.00x
🐘 Postgres	Nitro	1.066s (-44.7% 🟢)	1.576s (-25.0% 🟢)	0.510s	58	1.02x
💻 Local	Nitro	1.185s (-61.0% 🟢)	2.006s (-46.6% 🟢)	0.821s	45	1.13x
💻 Local	Express	1.191s (-60.5% 🟢)	2.005s (-44.1% 🟢)	0.815s	45	1.14x

workflow with 50 sequential data payload steps (10KB)

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Express	2.030s (-49.1% 🟢)	2.427s (-44.4% 🟢)	0.397s	50	1.00x
🐘 Postgres	Nitro	2.128s (-48.2% 🟢)	2.660s (-42.2% 🟢)	0.532s	46	1.05x
💻 Local	Nitro	2.704s (-70.9% 🟢)	3.032s (-69.7% 🟢)	0.328s	40	1.33x
💻 Local	Express	2.719s (-70.5% 🟢)	3.057s (-69.5% 🟢)	0.338s	40	1.34x

workflow with 10 concurrent data payload steps (10KB)

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Express	0.163s (-42.4% 🟢)	1.005s (~)	0.843s	60	1.00x
🐘 Postgres	Nitro	0.178s (-37.0% 🟢)	1.006s (~)	0.827s	60	1.10x
💻 Local	Express	0.455s (-18.8% 🟢)	1.004s (~)	0.550s	60	2.80x
💻 Local	Nitro	0.469s (-22.4% 🟢)	1.004s (-1.7%)	0.535s	60	2.89x

workflow with 25 concurrent data payload steps (10KB)

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Express	0.286s (-43.9% 🟢)	1.006s (~)	0.720s	90	1.00x
🐘 Postgres	Nitro	0.295s (-40.5% 🟢)	1.006s (~)	0.711s	90	1.03x
💻 Local	Nitro	2.205s (-13.1% 🟢)	2.944s (-2.2%)	0.738s	31	7.71x
💻 Local	Express	2.226s (-11.4% 🟢)	2.943s (-2.2%)	0.717s	31	7.78x

workflow with 50 concurrent data payload steps (10KB)

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Express	0.582s (-28.9% 🟢)	1.007s (-1.0%)	0.425s	120	1.00x
🐘 Postgres	Nitro	0.634s (-19.7% 🟢)	1.006s (~)	0.372s	120	1.09x
💻 Local	Nitro	10.021s (-10.5% 🟢)	10.611s (-9.0% 🟢)	0.590s	12	17.21x
💻 Local	Express	10.042s (-10.3% 🟢)	10.694s (-10.4% 🟢)	0.652s	12	17.25x

Stream Benchmarks(includes TTFB metrics)

workflow with stream

💻 Local Development

World	Framework	Workflow Time	TTFB	Slurp	Wall Time	Overhead	Samples	vs Fastest
💻 Local	🥇 Express	1.132s (+468.5% 🔺)	2.005s (+99.6% 🔺)	0.011s (-13.2% 🟢)	2.018s (+98.2% 🔺)	0.886s	10	1.00x
💻 Local	Nitro	1.133s (+430.4% 🔺)	2.005s (+99.6% 🔺)	0.011s (-15.2% 🟢)	2.018s (+98.1% 🔺)	0.885s	10	1.00x
🐘 Postgres	Nitro	1.135s (+453.6% 🔺)	2.001s (+100.2% 🔺)	0.001s (-20.0% 🟢)	2.010s (+98.7% 🔺)	0.875s	10	1.00x
🐘 Postgres	Express	1.135s (+453.5% 🔺)	1.999s (+100.2% 🔺)	0.001s (-25.0% 🟢)	2.010s (+98.7% 🔺)	0.875s	10	1.00x

stream pipeline with 5 transform steps (1MB)

💻 Local Development

World	Framework	Workflow Time	TTFB	Slurp	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Express	1.492s (+136.8% 🔺)	2.002s (+98.9% 🔺)	0.003s (-30.4% 🟢)	2.023s (+97.8% 🔺)	0.531s	30	1.00x
🐘 Postgres	Nitro	1.519s (+143.4% 🔺)	2.004s (+99.0% 🔺)	0.004s (-2.5%)	2.027s (+98.2% 🔺)	0.507s	30	1.02x
💻 Local	Nitro	1.707s (+103.5% 🔺)	2.013s (+98.9% 🔺)	0.009s (-4.3%)	2.203s (+97.4% 🔺)	0.496s	28	1.14x
💻 Local	Express	1.923s (+154.0% 🔺)	2.011s (+95.4% 🔺)	0.009s (-8.2% 🟢)	2.422s (+132.9% 🔺)	0.499s	25	1.29x

10 parallel streams (1MB each)

💻 Local Development

World	Framework	Workflow Time	TTFB	Slurp	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Express	0.627s (-34.7% 🟢)	1.033s (-19.2% 🟢)	0.000s (-20.7% 🟢)	1.045s (-20.0% 🟢)	0.418s	58	1.00x
🐘 Postgres	Nitro	0.662s (-31.7% 🟢)	1.029s (-17.5% 🟢)	0.000s (-17.2% 🟢)	1.049s (-16.6% 🟢)	0.387s	58	1.05x
💻 Local	Express	1.365s (+11.5% 🔺)	2.015s (~)	0.000s (+40.0% 🔺)	2.018s (~)	0.652s	30	2.18x
💻 Local	Nitro	1.561s (+27.6% 🔺)	2.016s (~)	0.000s (+364.3% 🔺)	2.197s (+8.7% 🔺)	0.636s	28	2.49x

fan-out fan-in 10 streams (1MB each)

💻 Local Development

World	Framework	Workflow Time	TTFB	Slurp	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Express	1.228s (-30.7% 🟢)	1.869s (-14.2% 🟢)	0.000s (NaN%)	1.900s (-13.6% 🟢)	0.673s	32	1.00x
🐘 Postgres	Nitro	1.325s (-26.1% 🟢)	1.997s (-6.7% 🟢)	0.000s (+86.7% 🔺)	2.019s (-7.2% 🟢)	0.694s	30	1.08x
💻 Local	Express	3.158s (-8.9% 🟢)	4.028s (~)	0.001s (-33.3% 🟢)	4.030s (~)	0.873s	15	2.57x
💻 Local	Nitro	3.187s (-5.9% 🟢)	4.029s (~)	0.000s (-75.0% 🟢)	4.031s (~)	0.844s	15	2.60x

Summary

Fastest Framework by World

Winner determined by most benchmark wins

World	🥇 Fastest Framework	Wins
💻 Local	Express	12/21
🐘 Postgres	Express	17/21

Fastest World by Framework

Winner determined by most benchmark wins

Framework	🥇 Fastest World	Wins
Express	🐘 Postgres	16/21
Nitro	🐘 Postgres	17/21

Column Definitions

• Workflow Time: Runtime reported by workflow (completedAt - createdAt) - primary metric
• TTFB: Time to First Byte - time from workflow start until first stream byte received (stream benchmarks only)
• Slurp: Time from first byte to complete stream consumption (stream benchmarks only)
• Wall Time: Total testbench time (trigger workflow + poll for result)
• Overhead: Testbench overhead (Wall Time - Workflow Time)
• Samples: Number of benchmark iterations run
• vs Fastest: How much slower compared to the fastest configuration for this benchmark

Worlds:

• 💻 Local: In-memory filesystem world (local development)
• 🐘 Postgres: PostgreSQL database world (local development)
• ▲ Vercel: Vercel production/preview deployment
• 🌐 Turso: Community world (local development)
• 🌐 MongoDB: Community world (local development)
• 🌐 Redis: Community world (local development)
• 🌐 Jazz: Community world (local development)
• 🌐 Redis: Community world (local development)
• 🌐 Redis + BullMQ: Community world (local development)
• 🌐 Cloudflare: Community world (local development)
• 🌐 MySQL: Community world (local development)
• 🌐 Azure: Community world (local development)
• 🌐 NATS JetStream: Community world (local development)
• 🌐 Upstash: Community world (local development)

---

📋 View full workflow run

---

❌ Some benchmark jobs failed:

• Local: success
• Postgres: success
• Vercel: failure

Check the workflow run for details.

[github-actions]

🧪 E2E Test Results

❌ Some tests failed

Summary

	Passed	Failed	Skipped	Total
❌ ▲ Vercel Production	1196	4	219	1419
✅ 💻 Local Development	1587	0	219	1806
✅ 📦 Local Production	1458	0	219	1677
❌ 🐘 Local Postgres	1585	2	219	1806
✅ 🪟 Windows	129	0	0	129
✅ 📋 Other	727	0	176	903
Total	6682	6	1052	7740

❌ Failed Tests

▲ Vercel Production (4 failed)

astro (1 failed):

• AbortController abortFromStepWorkflow: step abort cancels an in-flight sibling step

hono (2 failed):

• outputStreamWorkflow no startIndex (reads all chunks)
• stepFunctionPassingWorkflow - step function references can be passed as arguments (without closure vars) | wrun_01KRYMPWTNKF0CE4BP27X7R946 | 🔍 observability

nextjs-webpack (1 failed):

• parallelSleepWorkflow | wrun_01KRYME5H0F7D8EH2AAF56T0Z2 | 🔍 observability

🐘 Local Postgres (2 failed)

nextjs-webpack-stable-lazy-discovery-disabled (1 failed):

• DurableAgent e2e core basic text response

nextjs-webpack-stable-lazy-discovery-enabled (1 failed):

• addTenWorkflow | wrun_01KRYMBT8QHFK9EZ9FDY03D9G0

Details by Category

❌ ▲ Vercel Production

App	Passed	Failed	Skipped
❌ astro	102	1	26
✅ example	103	0	26
✅ express	103	0	26
✅ fastify	103	0	26
❌ hono	101	2	26
✅ nextjs-turbopack	127	0	2
❌ nextjs-webpack	126	1	2
✅ nitro	103	0	26
✅ nuxt	103	0	26
✅ sveltekit	122	0	7
✅ vite	103	0	26

✅ 💻 Local Development

App	Passed	Failed	Skipped
✅ astro-stable	104	0	25
✅ express-stable	104	0	25
✅ fastify-stable	104	0	25
✅ hono-stable	104	0	25
✅ nextjs-turbopack-canary	110	0	19
✅ nextjs-turbopack-stable-lazy-discovery-disabled	129	0	0
✅ nextjs-turbopack-stable-lazy-discovery-enabled	129	0	0
✅ nextjs-webpack-canary	110	0	19
✅ nextjs-webpack-stable-lazy-discovery-disabled	129	0	0
✅ nextjs-webpack-stable-lazy-discovery-enabled	129	0	0
✅ nitro-stable	104	0	25
✅ nuxt-stable	104	0	25
✅ sveltekit-stable	123	0	6
✅ vite-stable	104	0	25

✅ 📦 Local Production

App	Passed	Failed	Skipped
✅ astro-stable	104	0	25
✅ express-stable	104	0	25
✅ fastify-stable	104	0	25
✅ hono-stable	104	0	25
✅ nextjs-turbopack-canary	110	0	19
✅ nextjs-turbopack-stable-lazy-discovery-disabled	129	0	0
✅ nextjs-webpack-canary	110	0	19
✅ nextjs-webpack-stable-lazy-discovery-disabled	129	0	0
✅ nextjs-webpack-stable-lazy-discovery-enabled	129	0	0
✅ nitro-stable	104	0	25
✅ nuxt-stable	104	0	25
✅ sveltekit-stable	123	0	6
✅ vite-stable	104	0	25

❌ 🐘 Local Postgres

App	Passed	Failed	Skipped
✅ astro-stable	104	0	25
✅ express-stable	104	0	25
✅ fastify-stable	104	0	25
✅ hono-stable	104	0	25
✅ nextjs-turbopack-canary	110	0	19
✅ nextjs-turbopack-stable-lazy-discovery-disabled	129	0	0
✅ nextjs-turbopack-stable-lazy-discovery-enabled	129	0	0
✅ nextjs-webpack-canary	110	0	19
❌ nextjs-webpack-stable-lazy-discovery-disabled	128	1	0
❌ nextjs-webpack-stable-lazy-discovery-enabled	128	1	0
✅ nitro-stable	104	0	25
✅ nuxt-stable	104	0	25
✅ sveltekit-stable	123	0	6
✅ vite-stable	104	0	25

✅ 🪟 Windows

App	Passed	Failed	Skipped
✅ nextjs-turbopack	129	0	0

✅ 📋 Other

App	Passed	Failed	Skipped
✅ e2e-local-dev-nest-stable	104	0	25
✅ e2e-local-dev-tanstack-start-	104	0	25
✅ e2e-local-postgres-nest-stable	104	0	25
✅ e2e-local-postgres-tanstack-start-	104	0	25
✅ e2e-local-prod-nest-stable	104	0	25
✅ e2e-local-prod-tanstack-start-	104	0	25
✅ e2e-vercel-prod-tanstack-start	103	0	26

---

📋 View full workflow run

---

❌ Some E2E test jobs failed:

• Vercel Prod: failure
• Local Dev: success
• Local Prod: failure
• Local Postgres: failure
• Windows: success

Check the workflow run for details.

[Copilot]

Pull request overview

Adds a new "Versioning" Foundations page (v4 and v5) explaining default run pinning to deployments, recovery via cancel + rerun, opting into newer code with deploymentId: "latest", and carrying serializable state/streams across child-run continuations. Existing references to deploymentId: "latest" and versioning behavior are linked back to the new page. Cookbook pages also pick up callouts about long-lived / multi-turn patterns and, on v5 only, are updated to reflect that start() can now be called directly inside "use workflow" functions (via #1491), while v4 retains the step-wrapped pattern.

Changes:

• New foundations/versioning.mdx for both v4 and v5, plus meta.json + index.mdx + linkbacks from vercel-world, start.mdx, and code-transform.mdx.
• v5 cookbook + migration guides updated to drop "use step" wrappers around start() in favor of calling start() directly from the workflow function; v4 keeps the step-wrapped pattern.
• Added "Versioning" callouts to scheduling, durable-agent, ai-sdk, chat-sdk, and sandbox cookbooks; fixed a few small wording/grammar issues in code-transform.mdx.

Reviewed changes

Copilot reviewed 32 out of 32 changed files in this pull request and generated no comments.

Show a summary per file

File	Description
docs/content/docs/v5/foundations/versioning.mdx	New Foundations page explaining run pinning, recovery, and explicit upgrade boundaries.
docs/content/docs/v4/foundations/versioning.mdx	v4 counterpart; uses step-wrapped continueDigest/continueExportOnLatest helpers.
docs/content/docs/v{4,5}/foundations/{meta.json,index.mdx}	Surface the new Versioning page in nav and the foundations index.
docs/content/docs/v{4,5}/deploying/world/vercel-world.mdx	Link the deployment-pinning section to the new Versioning page.
docs/content/docs/v{4,5}/how-it-works/code-transform.mdx	Link "atomic versioning" reference; fix "function assume" / "from being upgraded" wording.
docs/content/docs/v{4,5}/api-reference/workflow-api/start.mdx	Soften "Vercel-specific" to "Vercelism", mention possible version rename, link Versioning. v5 also notes native workflow-context start().
docs/content/docs/v5/foundations/starting-workflows.mdx	Note v5 can call start() from workflows; link Composition + Versioning.
docs/content/docs/v5/cookbook/common-patterns/workflow-composition.mdx	Replace step-wrapped child spawn with direct start() call.
docs/content/docs/v5/cookbook/advanced/child-workflows.mdx	Inline batched start() calls; rename spawnReportChunk → startReportChunk; update tips.
docs/content/docs/v{4,5}/cookbook/common-patterns/scheduling.mdx	Callout linking to Versioning for long-lived schedules.
docs/content/docs/v{4,5}/cookbook/agent-patterns/durable-agent.mdx	Callout linking to Versioning for multi-turn agents.
docs/content/docs/v{4,5}/cookbook/integrations/{ai-sdk,chat-sdk,sandbox}.mdx	Callouts linking to Versioning for long-lived sessions.
docs/content/docs/v4/cookbook/common-patterns/workflow-composition.mdx	Update deploymentId: "latest" callout wording + link to Versioning.
docs/content/docs/v4/cookbook/advanced/child-workflows.mdx	Link Versioning from the deploymentId: "latest" tip.
docs/content/docs/v4/changelog/eager-processing.mdx	Reword to "step-wrapped start() helper" to match v4 phrasing.
docs/content/docs/v5/migration-guides/{migrating-from-trigger-dev,migrating-from-temporal,migrating-from-inngest,migrating-from-aws-step-functions}.mdx	Update to v5 native workflow-context start(); only collect step retains "use step".

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[TooTallNate]

Comment-only review

The new versioning foundations page is a useful addition — the four sections (default pinning → recovery → self-upgrade via start({ deploymentId: "latest" }) → carrying serializable context forward) cover the model clearly, and the v4-vs-v5 differentiation (step-wrapped start() vs direct start() from workflow context, gated on whether PR #1491 has landed on the target branch) is correctly handled. The cross-links from cookbook patterns to the new foundations page tie everything together nicely.

Three observations — see inline.

Inline comments

1. Wording: this PR introduces "Vercelism" as a new term and uses it 6 times across the diff, including replacing existing "Vercel-specific feature" wording in two <Callout>s. "Vercelism" reads as informal slang and isn't part of the established docs vocabulary. The existing tone uses "Vercel-specific", "Vercel deployments", "Vercel World", "non-Vercel" — all precise. Suggested restoring the prior phrasing. This is the only thing I'd consider blocking.
2. Broken link in v4: versioning.mdx references /docs/foundations/cancellation in its related frontmatter, but that page is v5-only. Easy fix.
3. Missing changeset: convention for docs-only PRs is an empty ---\n--- file.

A few small things I noticed

• The pre-emptively check for a conflict with getHookByToken() example (line 96 of versioning.mdx) — nice that you noted the race window in a later doc (#2011 callout), but this page doesn't have that warning. Worth a cross-link or <Callout type="warn"> here too, since the recovery pattern uses both start() and deploymentId: "latest" together.
• The _idempotency / using syntax is great. Worth confirming the docs site's syntax highlighter renders it — I noticed #1988 recently fixed a homepage link lint regression, and the using syntax is TypeScript 5.2+ specific.
• dailyDigest example (line 137-152): the state parameter is DigestState but the workflow doesn't actually use state.userId after the first sendDigest call — pattern is fine, just noting if you wanted to make the example illustrate state evolution more vividly, having sendDigest read state.lastSentAt (e.g. to avoid sending duplicates) would make the carry-forward intent clearer.

Overall the page reads well — once the Vercelism wording is restored and the v4 link is fixed, this should be good to go.

[TooTallNate]

All three review comments addressed in 7eff96c — Vercelism wording reverted, v4 broken cancellation link fixed, empty changeset added. LGTM, approving.

[github-actions]

Backport PR opened against stable: #2014. Merge conflicts were resolved by AI — please review carefully.

[pranaygp]

Summary

Adds a new Foundations page explaining Workflow SDK versioning semantics for both v5 and v4 docs.

The page covers:

• default run pinning to the deployment that started the run
• canceling and rerunning affected in-flight runs on the latest deployment
• explicit self-upgrade patterns using deploymentId: "latest"
• carrying serializable context such as streams and abort signals across child runs

This also links existing versioning and deploymentId: "latest" mentions back to the new Foundations page, and calls out that deploymentId is currently a Vercelism that may become a more generic version concept in the World spec later.

Follow-up cookbook pass added targeted callouts for long-lived or multi-turn patterns:

• scheduled workflows
• DurableAgent
• AI SDK multi-turn sessions
• Chat SDK sessions
• Sandbox sessions

The docs now also distinguish v5 from v4 for child-run starts. Native start() from workflow context was added by e295bae41 / #1491, which is on main but not stable, so v5 examples encourage direct workflow-context start() and v4 keeps the step-wrapped pattern.

Validation

• git diff --check
• Parsed the updated v4/v5 Foundations meta.json files as JSON
• Verified v4/v5 copies of the touched cookbook pages match before the v5-only native-start pass
• Searched docs/content for stale v5 step-wrapped start() guidance and v4 direct-start guidance
• Confirmed origin/stable lacks the 'use step' directive in packages/core/src/runtime/start.ts, while main has it via e295bae41 / feat: allow start() to be called directly inside workflow functions #1491

Full docs lint/typecheck were not run because this checkout does not have node_modules installed, so biome, vitest, and changeset are unavailable locally.

[vercel]

The latest updates on your projects. Learn more about Vercel for GitHub.

Project	Deployment	Actions	Updated (UTC)
example-nextjs-workflow-turbopack	Ready	Preview, Comment	May 18, 2026 10:49pm
example-nextjs-workflow-webpack	Ready	Preview, Comment	May 18, 2026 10:49pm
example-workflow	Ready	Preview, Comment	May 18, 2026 10:49pm
workbench-astro-workflow	Ready	Preview, Comment	May 18, 2026 10:49pm
workbench-express-workflow	Ready	Preview, Comment	May 18, 2026 10:49pm
workbench-fastify-workflow	Ready	Preview, Comment	May 18, 2026 10:49pm
workbench-hono-workflow	Ready	Preview, Comment	May 18, 2026 10:49pm
workbench-nitro-workflow	Ready	Preview, Comment	May 18, 2026 10:49pm
workbench-nuxt-workflow	Ready	Preview, Comment	May 18, 2026 10:49pm
workbench-sveltekit-workflow	Ready	Preview, Comment	May 18, 2026 10:49pm
workbench-tanstack-start-workflow	Ready	Preview, Comment	May 18, 2026 10:49pm
workbench-vite-workflow	Ready	Preview, Comment	May 18, 2026 10:49pm
workflow-docs	Ready	Preview, ✅ 11 resolved, Open in v0	May 18, 2026 10:49pm
workflow-swc-playground	Ready	Preview, Comment	May 18, 2026 10:49pm
workflow-tarballs	Ready	Preview, Comment	May 18, 2026 10:49pm
workflow-web	Ready	Preview, Comment	May 18, 2026 10:49pm

[changeset-bot]

🦋 Changeset detected

Latest commit: 7eff96c

The changes in this PR will be included in the next version bump.

This PR includes changesets to release 0 packages

When changesets are added to this PR, you'll see the packages that this PR includes changesets for and the associated semver types

Not sure what this means? Click here to learn what changesets are.

Click here if you're a maintainer who wants to add another changeset to this PR

[github-actions]

📊 Benchmark Results

📈 Comparing against baseline from main branch. Green 🟢 = faster, Red 🔺 = slower.

workflow with no steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
💻 Local	🥇 Nitro	0.031s (-28.3% 🟢)	1.005s (~)	0.974s	10	1.00x
💻 Local	Express	0.036s (-19.6% 🟢)	1.005s (~)	0.970s	10	1.15x
🐘 Postgres	Express	0.045s (-21.9% 🟢)	1.011s (~)	0.966s	10	1.47x
🐘 Postgres	Nitro	0.046s (-51.3% 🟢)	1.013s (-2.9%)	0.967s	10	1.50x

workflow with 1 step

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
💻 Local	🥇 Express	1.070s (-4.9%)	2.006s (~)	0.936s	10	1.00x
💻 Local	Nitro	1.071s (-5.3% 🟢)	2.006s (~)	0.935s	10	1.00x
🐘 Postgres	Express	1.079s (-5.9% 🟢)	2.009s (~)	0.930s	10	1.01x
🐘 Postgres	Nitro	1.081s (-5.2% 🟢)	2.010s (~)	0.929s	10	1.01x

workflow with 10 sequential steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Nitro	10.396s (-4.4%)	11.018s (~)	0.622s	3	1.00x
🐘 Postgres	Express	10.402s (-5.1% 🟢)	11.016s (~)	0.614s	3	1.00x
💻 Local	Express	10.411s (-4.7%)	11.023s (~)	0.612s	3	1.00x
💻 Local	Nitro	10.412s (-4.9%)	11.022s (~)	0.610s	3	1.00x

workflow with 25 sequential steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
💻 Local	🥇 Express	13.432s (-10.3% 🟢)	14.027s (-6.7% 🟢)	0.595s	5	1.00x
🐘 Postgres	Express	13.435s (-7.9% 🟢)	14.017s (-6.7% 🟢)	0.581s	5	1.00x
🐘 Postgres	Nitro	13.484s (-7.6% 🟢)	14.017s (-6.7% 🟢)	0.533s	5	1.00x
💻 Local	Nitro	13.489s (-10.4% 🟢)	14.027s (-12.5% 🟢)	0.538s	5	1.00x

workflow with 50 sequential steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
💻 Local	🥇 Nitro	11.875s (-29.2% 🟢)	12.022s (-29.4% 🟢)	0.147s	8	1.00x
💻 Local	Express	11.890s (-28.4% 🟢)	12.022s (-29.4% 🟢)	0.132s	8	1.00x
🐘 Postgres	Express	11.941s (-14.7% 🟢)	12.391s (-15.1% 🟢)	0.450s	8	1.01x
🐘 Postgres	Nitro	11.944s (-14.5% 🟢)	12.266s (-14.3% 🟢)	0.321s	8	1.01x

Promise.all with 10 concurrent steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Express	1.136s (-9.8% 🟢)	2.007s (~)	0.870s	15	1.00x
🐘 Postgres	Nitro	1.137s (-10.8% 🟢)	2.007s (~)	0.870s	15	1.00x
💻 Local	Nitro	1.172s (-28.2% 🟢)	2.006s (-3.3%)	0.834s	15	1.03x
💻 Local	Express	1.183s (-20.6% 🟢)	2.006s (~)	0.823s	15	1.04x

Promise.all with 25 concurrent steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Nitro	1.192s (-49.3% 🟢)	2.007s (-33.3% 🟢)	0.815s	15	1.00x
🐘 Postgres	Express	1.224s (-48.2% 🟢)	2.007s (-33.3% 🟢)	0.783s	15	1.03x
💻 Local	Express	1.703s (-42.3% 🟢)	2.005s (-41.9% 🟢)	0.303s	15	1.43x
💻 Local	Nitro	1.704s (-45.8% 🟢)	2.005s (-48.4% 🟢)	0.302s	15	1.43x

Promise.all with 50 concurrent steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Nitro	1.305s (-62.5% 🟢)	2.008s (-49.9% 🟢)	0.702s	15	1.00x
🐘 Postgres	Express	1.324s (-62.0% 🟢)	2.008s (-49.9% 🟢)	0.684s	15	1.01x
💻 Local	Express	4.638s (-44.4% 🟢)	5.179s (-42.6% 🟢)	0.541s	6	3.55x
💻 Local	Nitro	4.910s (-41.2% 🟢)	5.511s (-38.9% 🟢)	0.601s	6	3.76x

Promise.race with 10 concurrent steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Express	1.155s (-8.1% 🟢)	2.009s (~)	0.854s	15	1.00x
🐘 Postgres	Nitro	1.163s (-7.5% 🟢)	2.010s (~)	0.847s	15	1.01x
💻 Local	Express	1.386s (-26.8% 🟢)	2.006s (-15.2% 🟢)	0.620s	15	1.20x
💻 Local	Nitro	1.388s (-25.6% 🟢)	2.006s (-14.3% 🟢)	0.618s	15	1.20x

Promise.race with 25 concurrent steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Express	1.199s (-48.8% 🟢)	2.008s (-33.3% 🟢)	0.810s	15	1.00x
🐘 Postgres	Nitro	1.199s (-48.7% 🟢)	2.007s (-33.3% 🟢)	0.808s	15	1.00x
💻 Local	Express	1.828s (-41.6% 🟢)	2.222s (-40.9% 🟢)	0.394s	14	1.53x
💻 Local	Nitro	1.935s (-36.9% 🟢)	2.316s (-40.4% 🟢)	0.381s	13	1.61x

Promise.race with 50 concurrent steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Express	1.305s (-62.7% 🟢)	2.008s (-49.9% 🟢)	0.702s	15	1.00x
🐘 Postgres	Nitro	1.309s (-62.4% 🟢)	2.009s (-49.9% 🟢)	0.700s	15	1.00x
💻 Local	Express	5.446s (-38.1% 🟢)	5.849s (-36.9% 🟢)	0.403s	6	4.17x
💻 Local	Nitro	5.557s (-39.2% 🟢)	6.179s (-38.4% 🟢)	0.622s	6	4.26x

workflow with 10 sequential data payload steps (10KB)

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Express	0.418s (-50.1% 🟢)	1.006s (-1.6%)	0.588s	60	1.00x
🐘 Postgres	Nitro	0.459s (-44.0% 🟢)	1.006s (~)	0.547s	60	1.10x
💻 Local	Nitro	0.468s (-52.2% 🟢)	1.004s (-8.2% 🟢)	0.535s	60	1.12x
💻 Local	Express	0.469s (-52.4% 🟢)	1.004s (-6.7% 🟢)	0.535s	60	1.12x

workflow with 25 sequential data payload steps (10KB)

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Express	1.047s (-47.0% 🟢)	1.705s (-24.5% 🟢)	0.659s	53	1.00x
🐘 Postgres	Nitro	1.066s (-44.7% 🟢)	1.576s (-25.0% 🟢)	0.510s	58	1.02x
💻 Local	Nitro	1.185s (-61.0% 🟢)	2.006s (-46.6% 🟢)	0.821s	45	1.13x
💻 Local	Express	1.191s (-60.5% 🟢)	2.005s (-44.1% 🟢)	0.815s	45	1.14x

workflow with 50 sequential data payload steps (10KB)

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Express	2.030s (-49.1% 🟢)	2.427s (-44.4% 🟢)	0.397s	50	1.00x
🐘 Postgres	Nitro	2.128s (-48.2% 🟢)	2.660s (-42.2% 🟢)	0.532s	46	1.05x
💻 Local	Nitro	2.704s (-70.9% 🟢)	3.032s (-69.7% 🟢)	0.328s	40	1.33x
💻 Local	Express	2.719s (-70.5% 🟢)	3.057s (-69.5% 🟢)	0.338s	40	1.34x

workflow with 10 concurrent data payload steps (10KB)

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Express	0.163s (-42.4% 🟢)	1.005s (~)	0.843s	60	1.00x
🐘 Postgres	Nitro	0.178s (-37.0% 🟢)	1.006s (~)	0.827s	60	1.10x
💻 Local	Express	0.455s (-18.8% 🟢)	1.004s (~)	0.550s	60	2.80x
💻 Local	Nitro	0.469s (-22.4% 🟢)	1.004s (-1.7%)	0.535s	60	2.89x

workflow with 25 concurrent data payload steps (10KB)

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Express	0.286s (-43.9% 🟢)	1.006s (~)	0.720s	90	1.00x
🐘 Postgres	Nitro	0.295s (-40.5% 🟢)	1.006s (~)	0.711s	90	1.03x
💻 Local	Nitro	2.205s (-13.1% 🟢)	2.944s (-2.2%)	0.738s	31	7.71x
💻 Local	Express	2.226s (-11.4% 🟢)	2.943s (-2.2%)	0.717s	31	7.78x

workflow with 50 concurrent data payload steps (10KB)

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Express	0.582s (-28.9% 🟢)	1.007s (-1.0%)	0.425s	120	1.00x
🐘 Postgres	Nitro	0.634s (-19.7% 🟢)	1.006s (~)	0.372s	120	1.09x
💻 Local	Nitro	10.021s (-10.5% 🟢)	10.611s (-9.0% 🟢)	0.590s	12	17.21x
💻 Local	Express	10.042s (-10.3% 🟢)	10.694s (-10.4% 🟢)	0.652s	12	17.25x

Stream Benchmarks(includes TTFB metrics)

workflow with stream

💻 Local Development

World	Framework	Workflow Time	TTFB	Slurp	Wall Time	Overhead	Samples	vs Fastest
💻 Local	🥇 Express	1.132s (+468.5% 🔺)	2.005s (+99.6% 🔺)	0.011s (-13.2% 🟢)	2.018s (+98.2% 🔺)	0.886s	10	1.00x
💻 Local	Nitro	1.133s (+430.4% 🔺)	2.005s (+99.6% 🔺)	0.011s (-15.2% 🟢)	2.018s (+98.1% 🔺)	0.885s	10	1.00x
🐘 Postgres	Nitro	1.135s (+453.6% 🔺)	2.001s (+100.2% 🔺)	0.001s (-20.0% 🟢)	2.010s (+98.7% 🔺)	0.875s	10	1.00x
🐘 Postgres	Express	1.135s (+453.5% 🔺)	1.999s (+100.2% 🔺)	0.001s (-25.0% 🟢)	2.010s (+98.7% 🔺)	0.875s	10	1.00x

stream pipeline with 5 transform steps (1MB)

💻 Local Development

World	Framework	Workflow Time	TTFB	Slurp	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Express	1.492s (+136.8% 🔺)	2.002s (+98.9% 🔺)	0.003s (-30.4% 🟢)	2.023s (+97.8% 🔺)	0.531s	30	1.00x
🐘 Postgres	Nitro	1.519s (+143.4% 🔺)	2.004s (+99.0% 🔺)	0.004s (-2.5%)	2.027s (+98.2% 🔺)	0.507s	30	1.02x
💻 Local	Nitro	1.707s (+103.5% 🔺)	2.013s (+98.9% 🔺)	0.009s (-4.3%)	2.203s (+97.4% 🔺)	0.496s	28	1.14x
💻 Local	Express	1.923s (+154.0% 🔺)	2.011s (+95.4% 🔺)	0.009s (-8.2% 🟢)	2.422s (+132.9% 🔺)	0.499s	25	1.29x

10 parallel streams (1MB each)

💻 Local Development

World	Framework	Workflow Time	TTFB	Slurp	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Express	0.627s (-34.7% 🟢)	1.033s (-19.2% 🟢)	0.000s (-20.7% 🟢)	1.045s (-20.0% 🟢)	0.418s	58	1.00x
🐘 Postgres	Nitro	0.662s (-31.7% 🟢)	1.029s (-17.5% 🟢)	0.000s (-17.2% 🟢)	1.049s (-16.6% 🟢)	0.387s	58	1.05x
💻 Local	Express	1.365s (+11.5% 🔺)	2.015s (~)	0.000s (+40.0% 🔺)	2.018s (~)	0.652s	30	2.18x
💻 Local	Nitro	1.561s (+27.6% 🔺)	2.016s (~)	0.000s (+364.3% 🔺)	2.197s (+8.7% 🔺)	0.636s	28	2.49x

fan-out fan-in 10 streams (1MB each)

💻 Local Development

World	Framework	Workflow Time	TTFB	Slurp	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Express	1.228s (-30.7% 🟢)	1.869s (-14.2% 🟢)	0.000s (NaN%)	1.900s (-13.6% 🟢)	0.673s	32	1.00x
🐘 Postgres	Nitro	1.325s (-26.1% 🟢)	1.997s (-6.7% 🟢)	0.000s (+86.7% 🔺)	2.019s (-7.2% 🟢)	0.694s	30	1.08x
💻 Local	Express	3.158s (-8.9% 🟢)	4.028s (~)	0.001s (-33.3% 🟢)	4.030s (~)	0.873s	15	2.57x
💻 Local	Nitro	3.187s (-5.9% 🟢)	4.029s (~)	0.000s (-75.0% 🟢)	4.031s (~)	0.844s	15	2.60x

Summary

Fastest Framework by World

Winner determined by most benchmark wins

World	🥇 Fastest Framework	Wins
💻 Local	Express	12/21
🐘 Postgres	Express	17/21

Fastest World by Framework

Winner determined by most benchmark wins

Framework	🥇 Fastest World	Wins
Express	🐘 Postgres	16/21
Nitro	🐘 Postgres	17/21

Column Definitions

• Workflow Time: Runtime reported by workflow (completedAt - createdAt) - primary metric
• TTFB: Time to First Byte - time from workflow start until first stream byte received (stream benchmarks only)
• Slurp: Time from first byte to complete stream consumption (stream benchmarks only)
• Wall Time: Total testbench time (trigger workflow + poll for result)
• Overhead: Testbench overhead (Wall Time - Workflow Time)
• Samples: Number of benchmark iterations run
• vs Fastest: How much slower compared to the fastest configuration for this benchmark

Worlds:

• 💻 Local: In-memory filesystem world (local development)
• 🐘 Postgres: PostgreSQL database world (local development)
• ▲ Vercel: Vercel production/preview deployment
• 🌐 Turso: Community world (local development)
• 🌐 MongoDB: Community world (local development)
• 🌐 Redis: Community world (local development)
• 🌐 Jazz: Community world (local development)
• 🌐 Redis: Community world (local development)
• 🌐 Redis + BullMQ: Community world (local development)
• 🌐 Cloudflare: Community world (local development)
• 🌐 MySQL: Community world (local development)
• 🌐 Azure: Community world (local development)
• 🌐 NATS JetStream: Community world (local development)
• 🌐 Upstash: Community world (local development)

---

📋 View full workflow run

---

❌ Some benchmark jobs failed:

• Local: success
• Postgres: success
• Vercel: failure

Check the workflow run for details.

[github-actions]

🧪 E2E Test Results

❌ Some tests failed

Summary

	Passed	Failed	Skipped	Total
❌ ▲ Vercel Production	1196	4	219	1419
✅ 💻 Local Development	1587	0	219	1806
✅ 📦 Local Production	1458	0	219	1677
❌ 🐘 Local Postgres	1585	2	219	1806
✅ 🪟 Windows	129	0	0	129
✅ 📋 Other	727	0	176	903
Total	6682	6	1052	7740

❌ Failed Tests

▲ Vercel Production (4 failed)

astro (1 failed):

• AbortController abortFromStepWorkflow: step abort cancels an in-flight sibling step

hono (2 failed):

• outputStreamWorkflow no startIndex (reads all chunks)
• stepFunctionPassingWorkflow - step function references can be passed as arguments (without closure vars) | wrun_01KRYMPWTNKF0CE4BP27X7R946 | 🔍 observability

nextjs-webpack (1 failed):

• parallelSleepWorkflow | wrun_01KRYME5H0F7D8EH2AAF56T0Z2 | 🔍 observability

🐘 Local Postgres (2 failed)

nextjs-webpack-stable-lazy-discovery-disabled (1 failed):

• DurableAgent e2e core basic text response

nextjs-webpack-stable-lazy-discovery-enabled (1 failed):

• addTenWorkflow | wrun_01KRYMBT8QHFK9EZ9FDY03D9G0

Details by Category

❌ ▲ Vercel Production

App	Passed	Failed	Skipped
❌ astro	102	1	26
✅ example	103	0	26
✅ express	103	0	26
✅ fastify	103	0	26
❌ hono	101	2	26
✅ nextjs-turbopack	127	0	2
❌ nextjs-webpack	126	1	2
✅ nitro	103	0	26
✅ nuxt	103	0	26
✅ sveltekit	122	0	7
✅ vite	103	0	26

✅ 💻 Local Development

App	Passed	Failed	Skipped
✅ astro-stable	104	0	25
✅ express-stable	104	0	25
✅ fastify-stable	104	0	25
✅ hono-stable	104	0	25
✅ nextjs-turbopack-canary	110	0	19
✅ nextjs-turbopack-stable-lazy-discovery-disabled	129	0	0
✅ nextjs-turbopack-stable-lazy-discovery-enabled	129	0	0
✅ nextjs-webpack-canary	110	0	19
✅ nextjs-webpack-stable-lazy-discovery-disabled	129	0	0
✅ nextjs-webpack-stable-lazy-discovery-enabled	129	0	0
✅ nitro-stable	104	0	25
✅ nuxt-stable	104	0	25
✅ sveltekit-stable	123	0	6
✅ vite-stable	104	0	25

✅ 📦 Local Production

App	Passed	Failed	Skipped
✅ astro-stable	104	0	25
✅ express-stable	104	0	25
✅ fastify-stable	104	0	25
✅ hono-stable	104	0	25
✅ nextjs-turbopack-canary	110	0	19
✅ nextjs-turbopack-stable-lazy-discovery-disabled	129	0	0
✅ nextjs-webpack-canary	110	0	19
✅ nextjs-webpack-stable-lazy-discovery-disabled	129	0	0
✅ nextjs-webpack-stable-lazy-discovery-enabled	129	0	0
✅ nitro-stable	104	0	25
✅ nuxt-stable	104	0	25
✅ sveltekit-stable	123	0	6
✅ vite-stable	104	0	25

❌ 🐘 Local Postgres

App	Passed	Failed	Skipped
✅ astro-stable	104	0	25
✅ express-stable	104	0	25
✅ fastify-stable	104	0	25
✅ hono-stable	104	0	25
✅ nextjs-turbopack-canary	110	0	19
✅ nextjs-turbopack-stable-lazy-discovery-disabled	129	0	0
✅ nextjs-turbopack-stable-lazy-discovery-enabled	129	0	0
✅ nextjs-webpack-canary	110	0	19
❌ nextjs-webpack-stable-lazy-discovery-disabled	128	1	0
❌ nextjs-webpack-stable-lazy-discovery-enabled	128	1	0
✅ nitro-stable	104	0	25
✅ nuxt-stable	104	0	25
✅ sveltekit-stable	123	0	6
✅ vite-stable	104	0	25

✅ 🪟 Windows

App	Passed	Failed	Skipped
✅ nextjs-turbopack	129	0	0

✅ 📋 Other

App	Passed	Failed	Skipped
✅ e2e-local-dev-nest-stable	104	0	25
✅ e2e-local-dev-tanstack-start-	104	0	25
✅ e2e-local-postgres-nest-stable	104	0	25
✅ e2e-local-postgres-tanstack-start-	104	0	25
✅ e2e-local-prod-nest-stable	104	0	25
✅ e2e-local-prod-tanstack-start-	104	0	25
✅ e2e-vercel-prod-tanstack-start	103	0	26

---

📋 View full workflow run

---

❌ Some E2E test jobs failed:

• Vercel Prod: failure
• Local Dev: success
• Local Prod: failure
• Local Postgres: failure
• Windows: success

Check the workflow run for details.

[Copilot]

Pull request overview

Adds a new "Versioning" Foundations page (v4 and v5) explaining default run pinning to deployments, recovery via cancel + rerun, opting into newer code with deploymentId: "latest", and carrying serializable state/streams across child-run continuations. Existing references to deploymentId: "latest" and versioning behavior are linked back to the new page. Cookbook pages also pick up callouts about long-lived / multi-turn patterns and, on v5 only, are updated to reflect that start() can now be called directly inside "use workflow" functions (via #1491), while v4 retains the step-wrapped pattern.

Changes:

• New foundations/versioning.mdx for both v4 and v5, plus meta.json + index.mdx + linkbacks from vercel-world, start.mdx, and code-transform.mdx.
• v5 cookbook + migration guides updated to drop "use step" wrappers around start() in favor of calling start() directly from the workflow function; v4 keeps the step-wrapped pattern.
• Added "Versioning" callouts to scheduling, durable-agent, ai-sdk, chat-sdk, and sandbox cookbooks; fixed a few small wording/grammar issues in code-transform.mdx.

Reviewed changes

Copilot reviewed 32 out of 32 changed files in this pull request and generated no comments.

Show a summary per file

File	Description
docs/content/docs/v5/foundations/versioning.mdx	New Foundations page explaining run pinning, recovery, and explicit upgrade boundaries.
docs/content/docs/v4/foundations/versioning.mdx	v4 counterpart; uses step-wrapped continueDigest/continueExportOnLatest helpers.
docs/content/docs/v{4,5}/foundations/{meta.json,index.mdx}	Surface the new Versioning page in nav and the foundations index.
docs/content/docs/v{4,5}/deploying/world/vercel-world.mdx	Link the deployment-pinning section to the new Versioning page.
docs/content/docs/v{4,5}/how-it-works/code-transform.mdx	Link "atomic versioning" reference; fix "function assume" / "from being upgraded" wording.
docs/content/docs/v{4,5}/api-reference/workflow-api/start.mdx	Soften "Vercel-specific" to "Vercelism", mention possible version rename, link Versioning. v5 also notes native workflow-context start().
docs/content/docs/v5/foundations/starting-workflows.mdx	Note v5 can call start() from workflows; link Composition + Versioning.
docs/content/docs/v5/cookbook/common-patterns/workflow-composition.mdx	Replace step-wrapped child spawn with direct start() call.
docs/content/docs/v5/cookbook/advanced/child-workflows.mdx	Inline batched start() calls; rename spawnReportChunk → startReportChunk; update tips.
docs/content/docs/v{4,5}/cookbook/common-patterns/scheduling.mdx	Callout linking to Versioning for long-lived schedules.
docs/content/docs/v{4,5}/cookbook/agent-patterns/durable-agent.mdx	Callout linking to Versioning for multi-turn agents.
docs/content/docs/v{4,5}/cookbook/integrations/{ai-sdk,chat-sdk,sandbox}.mdx	Callouts linking to Versioning for long-lived sessions.
docs/content/docs/v4/cookbook/common-patterns/workflow-composition.mdx	Update deploymentId: "latest" callout wording + link to Versioning.
docs/content/docs/v4/cookbook/advanced/child-workflows.mdx	Link Versioning from the deploymentId: "latest" tip.
docs/content/docs/v4/changelog/eager-processing.mdx	Reword to "step-wrapped start() helper" to match v4 phrasing.
docs/content/docs/v5/migration-guides/{migrating-from-trigger-dev,migrating-from-temporal,migrating-from-inngest,migrating-from-aws-step-functions}.mdx	Update to v5 native workflow-context start(); only collect step retains "use step".

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[TooTallNate]

Comment-only review

The new versioning foundations page is a useful addition — the four sections (default pinning → recovery → self-upgrade via start({ deploymentId: "latest" }) → carrying serializable context forward) cover the model clearly, and the v4-vs-v5 differentiation (step-wrapped start() vs direct start() from workflow context, gated on whether PR #1491 has landed on the target branch) is correctly handled. The cross-links from cookbook patterns to the new foundations page tie everything together nicely.

Three observations — see inline.

Inline comments

1. Wording: this PR introduces "Vercelism" as a new term and uses it 6 times across the diff, including replacing existing "Vercel-specific feature" wording in two <Callout>s. "Vercelism" reads as informal slang and isn't part of the established docs vocabulary. The existing tone uses "Vercel-specific", "Vercel deployments", "Vercel World", "non-Vercel" — all precise. Suggested restoring the prior phrasing. This is the only thing I'd consider blocking.
2. Broken link in v4: versioning.mdx references /docs/foundations/cancellation in its related frontmatter, but that page is v5-only. Easy fix.
3. Missing changeset: convention for docs-only PRs is an empty ---\n--- file.

A few small things I noticed

• The pre-emptively check for a conflict with getHookByToken() example (line 96 of versioning.mdx) — nice that you noted the race window in a later doc (#2011 callout), but this page doesn't have that warning. Worth a cross-link or <Callout type="warn"> here too, since the recovery pattern uses both start() and deploymentId: "latest" together.
• The _idempotency / using syntax is great. Worth confirming the docs site's syntax highlighter renders it — I noticed #1988 recently fixed a homepage link lint regression, and the using syntax is TypeScript 5.2+ specific.
• dailyDigest example (line 137-152): the state parameter is DigestState but the workflow doesn't actually use state.userId after the first sendDigest call — pattern is fine, just noting if you wanted to make the example illustrate state evolution more vividly, having sendDigest read state.lastSentAt (e.g. to avoid sending duplicates) would make the carry-forward intent clearer.

Overall the page reads well — once the Vercelism wording is restored and the v4 link is fixed, this should be good to go.

[TooTallNate]

All three review comments addressed in 7eff96c — Vercelism wording reverted, v4 broken cancellation link fixed, empty changeset added. LGTM, approving.

[github-actions]

Backport PR opened against stable: #2014. Merge conflicts were resolved by AI — please review carefully.

[pranaygp]

Summary

Adds a new Foundations page explaining Workflow SDK versioning semantics for both v5 and v4 docs.

The page covers:

• default run pinning to the deployment that started the run
• canceling and rerunning affected in-flight runs on the latest deployment
• explicit self-upgrade patterns using deploymentId: "latest"
• carrying serializable context such as streams and abort signals across child runs

This also links existing versioning and deploymentId: "latest" mentions back to the new Foundations page, and calls out that deploymentId is currently a Vercelism that may become a more generic version concept in the World spec later.

Follow-up cookbook pass added targeted callouts for long-lived or multi-turn patterns:

• scheduled workflows
• DurableAgent
• AI SDK multi-turn sessions
• Chat SDK sessions
• Sandbox sessions

The docs now also distinguish v5 from v4 for child-run starts. Native start() from workflow context was added by e295bae41 / #1491, which is on main but not stable, so v5 examples encourage direct workflow-context start() and v4 keeps the step-wrapped pattern.

Validation

• git diff --check
• Parsed the updated v4/v5 Foundations meta.json files as JSON
• Verified v4/v5 copies of the touched cookbook pages match before the v5-only native-start pass
• Searched docs/content for stale v5 step-wrapped start() guidance and v4 direct-start guidance
• Confirmed origin/stable lacks the 'use step' directive in packages/core/src/runtime/start.ts, while main has it via e295bae41 / feat: allow start() to be called directly inside workflow functions #1491

Full docs lint/typecheck were not run because this checkout does not have node_modules installed, so biome, vitest, and changeset are unavailable locally.

[vercel]

The latest updates on your projects. Learn more about Vercel for GitHub.

Project	Deployment	Actions	Updated (UTC)
example-nextjs-workflow-turbopack	Ready	Preview, Comment	May 18, 2026 10:49pm
example-nextjs-workflow-webpack	Ready	Preview, Comment	May 18, 2026 10:49pm
example-workflow	Ready	Preview, Comment	May 18, 2026 10:49pm
workbench-astro-workflow	Ready	Preview, Comment	May 18, 2026 10:49pm
workbench-express-workflow	Ready	Preview, Comment	May 18, 2026 10:49pm
workbench-fastify-workflow	Ready	Preview, Comment	May 18, 2026 10:49pm
workbench-hono-workflow	Ready	Preview, Comment	May 18, 2026 10:49pm
workbench-nitro-workflow	Ready	Preview, Comment	May 18, 2026 10:49pm
workbench-nuxt-workflow	Ready	Preview, Comment	May 18, 2026 10:49pm
workbench-sveltekit-workflow	Ready	Preview, Comment	May 18, 2026 10:49pm
workbench-tanstack-start-workflow	Ready	Preview, Comment	May 18, 2026 10:49pm
workbench-vite-workflow	Ready	Preview, Comment	May 18, 2026 10:49pm
workflow-docs	Ready	Preview, ✅ 11 resolved, Open in v0	May 18, 2026 10:49pm
workflow-swc-playground	Ready	Preview, Comment	May 18, 2026 10:49pm
workflow-tarballs	Ready	Preview, Comment	May 18, 2026 10:49pm
workflow-web	Ready	Preview, Comment	May 18, 2026 10:49pm

[changeset-bot]

🦋 Changeset detected

Latest commit: 7eff96c

The changes in this PR will be included in the next version bump.

This PR includes changesets to release 0 packages

When changesets are added to this PR, you'll see the packages that this PR includes changesets for and the associated semver types

Not sure what this means? Click here to learn what changesets are.

Click here if you're a maintainer who wants to add another changeset to this PR

[github-actions]

📊 Benchmark Results

📈 Comparing against baseline from main branch. Green 🟢 = faster, Red 🔺 = slower.

workflow with no steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
💻 Local	🥇 Nitro	0.031s (-28.3% 🟢)	1.005s (~)	0.974s	10	1.00x
💻 Local	Express	0.036s (-19.6% 🟢)	1.005s (~)	0.970s	10	1.15x
🐘 Postgres	Express	0.045s (-21.9% 🟢)	1.011s (~)	0.966s	10	1.47x
🐘 Postgres	Nitro	0.046s (-51.3% 🟢)	1.013s (-2.9%)	0.967s	10	1.50x

workflow with 1 step

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
💻 Local	🥇 Express	1.070s (-4.9%)	2.006s (~)	0.936s	10	1.00x
💻 Local	Nitro	1.071s (-5.3% 🟢)	2.006s (~)	0.935s	10	1.00x
🐘 Postgres	Express	1.079s (-5.9% 🟢)	2.009s (~)	0.930s	10	1.01x
🐘 Postgres	Nitro	1.081s (-5.2% 🟢)	2.010s (~)	0.929s	10	1.01x

workflow with 10 sequential steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Nitro	10.396s (-4.4%)	11.018s (~)	0.622s	3	1.00x
🐘 Postgres	Express	10.402s (-5.1% 🟢)	11.016s (~)	0.614s	3	1.00x
💻 Local	Express	10.411s (-4.7%)	11.023s (~)	0.612s	3	1.00x
💻 Local	Nitro	10.412s (-4.9%)	11.022s (~)	0.610s	3	1.00x

workflow with 25 sequential steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
💻 Local	🥇 Express	13.432s (-10.3% 🟢)	14.027s (-6.7% 🟢)	0.595s	5	1.00x
🐘 Postgres	Express	13.435s (-7.9% 🟢)	14.017s (-6.7% 🟢)	0.581s	5	1.00x
🐘 Postgres	Nitro	13.484s (-7.6% 🟢)	14.017s (-6.7% 🟢)	0.533s	5	1.00x
💻 Local	Nitro	13.489s (-10.4% 🟢)	14.027s (-12.5% 🟢)	0.538s	5	1.00x

workflow with 50 sequential steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
💻 Local	🥇 Nitro	11.875s (-29.2% 🟢)	12.022s (-29.4% 🟢)	0.147s	8	1.00x
💻 Local	Express	11.890s (-28.4% 🟢)	12.022s (-29.4% 🟢)	0.132s	8	1.00x
🐘 Postgres	Express	11.941s (-14.7% 🟢)	12.391s (-15.1% 🟢)	0.450s	8	1.01x
🐘 Postgres	Nitro	11.944s (-14.5% 🟢)	12.266s (-14.3% 🟢)	0.321s	8	1.01x

Promise.all with 10 concurrent steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Express	1.136s (-9.8% 🟢)	2.007s (~)	0.870s	15	1.00x
🐘 Postgres	Nitro	1.137s (-10.8% 🟢)	2.007s (~)	0.870s	15	1.00x
💻 Local	Nitro	1.172s (-28.2% 🟢)	2.006s (-3.3%)	0.834s	15	1.03x
💻 Local	Express	1.183s (-20.6% 🟢)	2.006s (~)	0.823s	15	1.04x

Promise.all with 25 concurrent steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Nitro	1.192s (-49.3% 🟢)	2.007s (-33.3% 🟢)	0.815s	15	1.00x
🐘 Postgres	Express	1.224s (-48.2% 🟢)	2.007s (-33.3% 🟢)	0.783s	15	1.03x
💻 Local	Express	1.703s (-42.3% 🟢)	2.005s (-41.9% 🟢)	0.303s	15	1.43x
💻 Local	Nitro	1.704s (-45.8% 🟢)	2.005s (-48.4% 🟢)	0.302s	15	1.43x

Promise.all with 50 concurrent steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Nitro	1.305s (-62.5% 🟢)	2.008s (-49.9% 🟢)	0.702s	15	1.00x
🐘 Postgres	Express	1.324s (-62.0% 🟢)	2.008s (-49.9% 🟢)	0.684s	15	1.01x
💻 Local	Express	4.638s (-44.4% 🟢)	5.179s (-42.6% 🟢)	0.541s	6	3.55x
💻 Local	Nitro	4.910s (-41.2% 🟢)	5.511s (-38.9% 🟢)	0.601s	6	3.76x

Promise.race with 10 concurrent steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Express	1.155s (-8.1% 🟢)	2.009s (~)	0.854s	15	1.00x
🐘 Postgres	Nitro	1.163s (-7.5% 🟢)	2.010s (~)	0.847s	15	1.01x
💻 Local	Express	1.386s (-26.8% 🟢)	2.006s (-15.2% 🟢)	0.620s	15	1.20x
💻 Local	Nitro	1.388s (-25.6% 🟢)	2.006s (-14.3% 🟢)	0.618s	15	1.20x

Promise.race with 25 concurrent steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Express	1.199s (-48.8% 🟢)	2.008s (-33.3% 🟢)	0.810s	15	1.00x
🐘 Postgres	Nitro	1.199s (-48.7% 🟢)	2.007s (-33.3% 🟢)	0.808s	15	1.00x
💻 Local	Express	1.828s (-41.6% 🟢)	2.222s (-40.9% 🟢)	0.394s	14	1.53x
💻 Local	Nitro	1.935s (-36.9% 🟢)	2.316s (-40.4% 🟢)	0.381s	13	1.61x

Promise.race with 50 concurrent steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Express	1.305s (-62.7% 🟢)	2.008s (-49.9% 🟢)	0.702s	15	1.00x
🐘 Postgres	Nitro	1.309s (-62.4% 🟢)	2.009s (-49.9% 🟢)	0.700s	15	1.00x
💻 Local	Express	5.446s (-38.1% 🟢)	5.849s (-36.9% 🟢)	0.403s	6	4.17x
💻 Local	Nitro	5.557s (-39.2% 🟢)	6.179s (-38.4% 🟢)	0.622s	6	4.26x

workflow with 10 sequential data payload steps (10KB)

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Express	0.418s (-50.1% 🟢)	1.006s (-1.6%)	0.588s	60	1.00x
🐘 Postgres	Nitro	0.459s (-44.0% 🟢)	1.006s (~)	0.547s	60	1.10x
💻 Local	Nitro	0.468s (-52.2% 🟢)	1.004s (-8.2% 🟢)	0.535s	60	1.12x
💻 Local	Express	0.469s (-52.4% 🟢)	1.004s (-6.7% 🟢)	0.535s	60	1.12x

workflow with 25 sequential data payload steps (10KB)

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Express	1.047s (-47.0% 🟢)	1.705s (-24.5% 🟢)	0.659s	53	1.00x
🐘 Postgres	Nitro	1.066s (-44.7% 🟢)	1.576s (-25.0% 🟢)	0.510s	58	1.02x
💻 Local	Nitro	1.185s (-61.0% 🟢)	2.006s (-46.6% 🟢)	0.821s	45	1.13x
💻 Local	Express	1.191s (-60.5% 🟢)	2.005s (-44.1% 🟢)	0.815s	45	1.14x

workflow with 50 sequential data payload steps (10KB)

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Express	2.030s (-49.1% 🟢)	2.427s (-44.4% 🟢)	0.397s	50	1.00x
🐘 Postgres	Nitro	2.128s (-48.2% 🟢)	2.660s (-42.2% 🟢)	0.532s	46	1.05x
💻 Local	Nitro	2.704s (-70.9% 🟢)	3.032s (-69.7% 🟢)	0.328s	40	1.33x
💻 Local	Express	2.719s (-70.5% 🟢)	3.057s (-69.5% 🟢)	0.338s	40	1.34x

workflow with 10 concurrent data payload steps (10KB)

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Express	0.163s (-42.4% 🟢)	1.005s (~)	0.843s	60	1.00x
🐘 Postgres	Nitro	0.178s (-37.0% 🟢)	1.006s (~)	0.827s	60	1.10x
💻 Local	Express	0.455s (-18.8% 🟢)	1.004s (~)	0.550s	60	2.80x
💻 Local	Nitro	0.469s (-22.4% 🟢)	1.004s (-1.7%)	0.535s	60	2.89x

workflow with 25 concurrent data payload steps (10KB)

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Express	0.286s (-43.9% 🟢)	1.006s (~)	0.720s	90	1.00x
🐘 Postgres	Nitro	0.295s (-40.5% 🟢)	1.006s (~)	0.711s	90	1.03x
💻 Local	Nitro	2.205s (-13.1% 🟢)	2.944s (-2.2%)	0.738s	31	7.71x
💻 Local	Express	2.226s (-11.4% 🟢)	2.943s (-2.2%)	0.717s	31	7.78x

workflow with 50 concurrent data payload steps (10KB)

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Express	0.582s (-28.9% 🟢)	1.007s (-1.0%)	0.425s	120	1.00x
🐘 Postgres	Nitro	0.634s (-19.7% 🟢)	1.006s (~)	0.372s	120	1.09x
💻 Local	Nitro	10.021s (-10.5% 🟢)	10.611s (-9.0% 🟢)	0.590s	12	17.21x
💻 Local	Express	10.042s (-10.3% 🟢)	10.694s (-10.4% 🟢)	0.652s	12	17.25x

Stream Benchmarks(includes TTFB metrics)

workflow with stream

💻 Local Development

World	Framework	Workflow Time	TTFB	Slurp	Wall Time	Overhead	Samples	vs Fastest
💻 Local	🥇 Express	1.132s (+468.5% 🔺)	2.005s (+99.6% 🔺)	0.011s (-13.2% 🟢)	2.018s (+98.2% 🔺)	0.886s	10	1.00x
💻 Local	Nitro	1.133s (+430.4% 🔺)	2.005s (+99.6% 🔺)	0.011s (-15.2% 🟢)	2.018s (+98.1% 🔺)	0.885s	10	1.00x
🐘 Postgres	Nitro	1.135s (+453.6% 🔺)	2.001s (+100.2% 🔺)	0.001s (-20.0% 🟢)	2.010s (+98.7% 🔺)	0.875s	10	1.00x
🐘 Postgres	Express	1.135s (+453.5% 🔺)	1.999s (+100.2% 🔺)	0.001s (-25.0% 🟢)	2.010s (+98.7% 🔺)	0.875s	10	1.00x

stream pipeline with 5 transform steps (1MB)

💻 Local Development

World	Framework	Workflow Time	TTFB	Slurp	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Express	1.492s (+136.8% 🔺)	2.002s (+98.9% 🔺)	0.003s (-30.4% 🟢)	2.023s (+97.8% 🔺)	0.531s	30	1.00x
🐘 Postgres	Nitro	1.519s (+143.4% 🔺)	2.004s (+99.0% 🔺)	0.004s (-2.5%)	2.027s (+98.2% 🔺)	0.507s	30	1.02x
💻 Local	Nitro	1.707s (+103.5% 🔺)	2.013s (+98.9% 🔺)	0.009s (-4.3%)	2.203s (+97.4% 🔺)	0.496s	28	1.14x
💻 Local	Express	1.923s (+154.0% 🔺)	2.011s (+95.4% 🔺)	0.009s (-8.2% 🟢)	2.422s (+132.9% 🔺)	0.499s	25	1.29x

10 parallel streams (1MB each)

💻 Local Development

World	Framework	Workflow Time	TTFB	Slurp	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Express	0.627s (-34.7% 🟢)	1.033s (-19.2% 🟢)	0.000s (-20.7% 🟢)	1.045s (-20.0% 🟢)	0.418s	58	1.00x
🐘 Postgres	Nitro	0.662s (-31.7% 🟢)	1.029s (-17.5% 🟢)	0.000s (-17.2% 🟢)	1.049s (-16.6% 🟢)	0.387s	58	1.05x
💻 Local	Express	1.365s (+11.5% 🔺)	2.015s (~)	0.000s (+40.0% 🔺)	2.018s (~)	0.652s	30	2.18x
💻 Local	Nitro	1.561s (+27.6% 🔺)	2.016s (~)	0.000s (+364.3% 🔺)	2.197s (+8.7% 🔺)	0.636s	28	2.49x

fan-out fan-in 10 streams (1MB each)

💻 Local Development

World	Framework	Workflow Time	TTFB	Slurp	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Express	1.228s (-30.7% 🟢)	1.869s (-14.2% 🟢)	0.000s (NaN%)	1.900s (-13.6% 🟢)	0.673s	32	1.00x
🐘 Postgres	Nitro	1.325s (-26.1% 🟢)	1.997s (-6.7% 🟢)	0.000s (+86.7% 🔺)	2.019s (-7.2% 🟢)	0.694s	30	1.08x
💻 Local	Express	3.158s (-8.9% 🟢)	4.028s (~)	0.001s (-33.3% 🟢)	4.030s (~)	0.873s	15	2.57x
💻 Local	Nitro	3.187s (-5.9% 🟢)	4.029s (~)	0.000s (-75.0% 🟢)	4.031s (~)	0.844s	15	2.60x

Summary

Fastest Framework by World

Winner determined by most benchmark wins

World	🥇 Fastest Framework	Wins
💻 Local	Express	12/21
🐘 Postgres	Express	17/21

Fastest World by Framework

Winner determined by most benchmark wins

Framework	🥇 Fastest World	Wins
Express	🐘 Postgres	16/21
Nitro	🐘 Postgres	17/21

Column Definitions

• Workflow Time: Runtime reported by workflow (completedAt - createdAt) - primary metric
• TTFB: Time to First Byte - time from workflow start until first stream byte received (stream benchmarks only)
• Slurp: Time from first byte to complete stream consumption (stream benchmarks only)
• Wall Time: Total testbench time (trigger workflow + poll for result)
• Overhead: Testbench overhead (Wall Time - Workflow Time)
• Samples: Number of benchmark iterations run
• vs Fastest: How much slower compared to the fastest configuration for this benchmark

Worlds:

• 💻 Local: In-memory filesystem world (local development)
• 🐘 Postgres: PostgreSQL database world (local development)
• ▲ Vercel: Vercel production/preview deployment
• 🌐 Turso: Community world (local development)
• 🌐 MongoDB: Community world (local development)
• 🌐 Redis: Community world (local development)
• 🌐 Jazz: Community world (local development)
• 🌐 Redis: Community world (local development)
• 🌐 Redis + BullMQ: Community world (local development)
• 🌐 Cloudflare: Community world (local development)
• 🌐 MySQL: Community world (local development)
• 🌐 Azure: Community world (local development)
• 🌐 NATS JetStream: Community world (local development)
• 🌐 Upstash: Community world (local development)

---

📋 View full workflow run

---

❌ Some benchmark jobs failed:

• Local: success
• Postgres: success
• Vercel: failure

Check the workflow run for details.

[github-actions]

🧪 E2E Test Results

❌ Some tests failed

Summary

	Passed	Failed	Skipped	Total
❌ ▲ Vercel Production	1196	4	219	1419
✅ 💻 Local Development	1587	0	219	1806
✅ 📦 Local Production	1458	0	219	1677
❌ 🐘 Local Postgres	1585	2	219	1806
✅ 🪟 Windows	129	0	0	129
✅ 📋 Other	727	0	176	903
Total	6682	6	1052	7740

❌ Failed Tests

▲ Vercel Production (4 failed)

astro (1 failed):

• AbortController abortFromStepWorkflow: step abort cancels an in-flight sibling step

hono (2 failed):

• outputStreamWorkflow no startIndex (reads all chunks)
• stepFunctionPassingWorkflow - step function references can be passed as arguments (without closure vars) | wrun_01KRYMPWTNKF0CE4BP27X7R946 | 🔍 observability

nextjs-webpack (1 failed):

• parallelSleepWorkflow | wrun_01KRYME5H0F7D8EH2AAF56T0Z2 | 🔍 observability

🐘 Local Postgres (2 failed)

nextjs-webpack-stable-lazy-discovery-disabled (1 failed):

• DurableAgent e2e core basic text response

nextjs-webpack-stable-lazy-discovery-enabled (1 failed):

• addTenWorkflow | wrun_01KRYMBT8QHFK9EZ9FDY03D9G0

Details by Category

❌ ▲ Vercel Production

App	Passed	Failed	Skipped
❌ astro	102	1	26
✅ example	103	0	26
✅ express	103	0	26
✅ fastify	103	0	26
❌ hono	101	2	26
✅ nextjs-turbopack	127	0	2
❌ nextjs-webpack	126	1	2
✅ nitro	103	0	26
✅ nuxt	103	0	26
✅ sveltekit	122	0	7
✅ vite	103	0	26

✅ 💻 Local Development

App	Passed	Failed	Skipped
✅ astro-stable	104	0	25
✅ express-stable	104	0	25
✅ fastify-stable	104	0	25
✅ hono-stable	104	0	25
✅ nextjs-turbopack-canary	110	0	19
✅ nextjs-turbopack-stable-lazy-discovery-disabled	129	0	0
✅ nextjs-turbopack-stable-lazy-discovery-enabled	129	0	0
✅ nextjs-webpack-canary	110	0	19
✅ nextjs-webpack-stable-lazy-discovery-disabled	129	0	0
✅ nextjs-webpack-stable-lazy-discovery-enabled	129	0	0
✅ nitro-stable	104	0	25
✅ nuxt-stable	104	0	25
✅ sveltekit-stable	123	0	6
✅ vite-stable	104	0	25

✅ 📦 Local Production

App	Passed	Failed	Skipped
✅ astro-stable	104	0	25
✅ express-stable	104	0	25
✅ fastify-stable	104	0	25
✅ hono-stable	104	0	25
✅ nextjs-turbopack-canary	110	0	19
✅ nextjs-turbopack-stable-lazy-discovery-disabled	129	0	0
✅ nextjs-webpack-canary	110	0	19
✅ nextjs-webpack-stable-lazy-discovery-disabled	129	0	0
✅ nextjs-webpack-stable-lazy-discovery-enabled	129	0	0
✅ nitro-stable	104	0	25
✅ nuxt-stable	104	0	25
✅ sveltekit-stable	123	0	6
✅ vite-stable	104	0	25

❌ 🐘 Local Postgres

App	Passed	Failed	Skipped
✅ astro-stable	104	0	25
✅ express-stable	104	0	25
✅ fastify-stable	104	0	25
✅ hono-stable	104	0	25
✅ nextjs-turbopack-canary	110	0	19
✅ nextjs-turbopack-stable-lazy-discovery-disabled	129	0	0
✅ nextjs-turbopack-stable-lazy-discovery-enabled	129	0	0
✅ nextjs-webpack-canary	110	0	19
❌ nextjs-webpack-stable-lazy-discovery-disabled	128	1	0
❌ nextjs-webpack-stable-lazy-discovery-enabled	128	1	0
✅ nitro-stable	104	0	25
✅ nuxt-stable	104	0	25
✅ sveltekit-stable	123	0	6
✅ vite-stable	104	0	25

✅ 🪟 Windows

App	Passed	Failed	Skipped
✅ nextjs-turbopack	129	0	0

✅ 📋 Other

App	Passed	Failed	Skipped
✅ e2e-local-dev-nest-stable	104	0	25
✅ e2e-local-dev-tanstack-start-	104	0	25
✅ e2e-local-postgres-nest-stable	104	0	25
✅ e2e-local-postgres-tanstack-start-	104	0	25
✅ e2e-local-prod-nest-stable	104	0	25
✅ e2e-local-prod-tanstack-start-	104	0	25
✅ e2e-vercel-prod-tanstack-start	103	0	26

---

📋 View full workflow run

---

❌ Some E2E test jobs failed:

• Vercel Prod: failure
• Local Dev: success
• Local Prod: failure
• Local Postgres: failure
• Windows: success

Check the workflow run for details.

[Copilot]

Pull request overview

Adds a new "Versioning" Foundations page (v4 and v5) explaining default run pinning to deployments, recovery via cancel + rerun, opting into newer code with deploymentId: "latest", and carrying serializable state/streams across child-run continuations. Existing references to deploymentId: "latest" and versioning behavior are linked back to the new page. Cookbook pages also pick up callouts about long-lived / multi-turn patterns and, on v5 only, are updated to reflect that start() can now be called directly inside "use workflow" functions (via #1491), while v4 retains the step-wrapped pattern.

Changes:

• New foundations/versioning.mdx for both v4 and v5, plus meta.json + index.mdx + linkbacks from vercel-world, start.mdx, and code-transform.mdx.
• v5 cookbook + migration guides updated to drop "use step" wrappers around start() in favor of calling start() directly from the workflow function; v4 keeps the step-wrapped pattern.
• Added "Versioning" callouts to scheduling, durable-agent, ai-sdk, chat-sdk, and sandbox cookbooks; fixed a few small wording/grammar issues in code-transform.mdx.

Reviewed changes

Copilot reviewed 32 out of 32 changed files in this pull request and generated no comments.

Show a summary per file

File	Description
docs/content/docs/v5/foundations/versioning.mdx	New Foundations page explaining run pinning, recovery, and explicit upgrade boundaries.
docs/content/docs/v4/foundations/versioning.mdx	v4 counterpart; uses step-wrapped continueDigest/continueExportOnLatest helpers.
docs/content/docs/v{4,5}/foundations/{meta.json,index.mdx}	Surface the new Versioning page in nav and the foundations index.
docs/content/docs/v{4,5}/deploying/world/vercel-world.mdx	Link the deployment-pinning section to the new Versioning page.
docs/content/docs/v{4,5}/how-it-works/code-transform.mdx	Link "atomic versioning" reference; fix "function assume" / "from being upgraded" wording.
docs/content/docs/v{4,5}/api-reference/workflow-api/start.mdx	Soften "Vercel-specific" to "Vercelism", mention possible version rename, link Versioning. v5 also notes native workflow-context start().
docs/content/docs/v5/foundations/starting-workflows.mdx	Note v5 can call start() from workflows; link Composition + Versioning.
docs/content/docs/v5/cookbook/common-patterns/workflow-composition.mdx	Replace step-wrapped child spawn with direct start() call.
docs/content/docs/v5/cookbook/advanced/child-workflows.mdx	Inline batched start() calls; rename spawnReportChunk → startReportChunk; update tips.
docs/content/docs/v{4,5}/cookbook/common-patterns/scheduling.mdx	Callout linking to Versioning for long-lived schedules.
docs/content/docs/v{4,5}/cookbook/agent-patterns/durable-agent.mdx	Callout linking to Versioning for multi-turn agents.
docs/content/docs/v{4,5}/cookbook/integrations/{ai-sdk,chat-sdk,sandbox}.mdx	Callouts linking to Versioning for long-lived sessions.
docs/content/docs/v4/cookbook/common-patterns/workflow-composition.mdx	Update deploymentId: "latest" callout wording + link to Versioning.
docs/content/docs/v4/cookbook/advanced/child-workflows.mdx	Link Versioning from the deploymentId: "latest" tip.
docs/content/docs/v4/changelog/eager-processing.mdx	Reword to "step-wrapped start() helper" to match v4 phrasing.
docs/content/docs/v5/migration-guides/{migrating-from-trigger-dev,migrating-from-temporal,migrating-from-inngest,migrating-from-aws-step-functions}.mdx	Update to v5 native workflow-context start(); only collect step retains "use step".

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[TooTallNate]

Comment-only review

The new versioning foundations page is a useful addition — the four sections (default pinning → recovery → self-upgrade via start({ deploymentId: "latest" }) → carrying serializable context forward) cover the model clearly, and the v4-vs-v5 differentiation (step-wrapped start() vs direct start() from workflow context, gated on whether PR #1491 has landed on the target branch) is correctly handled. The cross-links from cookbook patterns to the new foundations page tie everything together nicely.

Three observations — see inline.

Inline comments

1. Wording: this PR introduces "Vercelism" as a new term and uses it 6 times across the diff, including replacing existing "Vercel-specific feature" wording in two <Callout>s. "Vercelism" reads as informal slang and isn't part of the established docs vocabulary. The existing tone uses "Vercel-specific", "Vercel deployments", "Vercel World", "non-Vercel" — all precise. Suggested restoring the prior phrasing. This is the only thing I'd consider blocking.
2. Broken link in v4: versioning.mdx references /docs/foundations/cancellation in its related frontmatter, but that page is v5-only. Easy fix.
3. Missing changeset: convention for docs-only PRs is an empty ---\n--- file.

A few small things I noticed

• The pre-emptively check for a conflict with getHookByToken() example (line 96 of versioning.mdx) — nice that you noted the race window in a later doc (#2011 callout), but this page doesn't have that warning. Worth a cross-link or <Callout type="warn"> here too, since the recovery pattern uses both start() and deploymentId: "latest" together.
• The _idempotency / using syntax is great. Worth confirming the docs site's syntax highlighter renders it — I noticed #1988 recently fixed a homepage link lint regression, and the using syntax is TypeScript 5.2+ specific.
• dailyDigest example (line 137-152): the state parameter is DigestState but the workflow doesn't actually use state.userId after the first sendDigest call — pattern is fine, just noting if you wanted to make the example illustrate state evolution more vividly, having sendDigest read state.lastSentAt (e.g. to avoid sending duplicates) would make the carry-forward intent clearer.

Overall the page reads well — once the Vercelism wording is restored and the v4 link is fixed, this should be good to go.

[TooTallNate]

All three review comments addressed in 7eff96c — Vercelism wording reverted, v4 broken cancellation link fixed, empty changeset added. LGTM, approving.

[github-actions]

Backport PR opened against stable: #2014. Merge conflicts were resolved by AI — please review carefully.

[pranaygp]

Summary

Adds a new Foundations page explaining Workflow SDK versioning semantics for both v5 and v4 docs.

The page covers:

• default run pinning to the deployment that started the run
• canceling and rerunning affected in-flight runs on the latest deployment
• explicit self-upgrade patterns using deploymentId: "latest"
• carrying serializable context such as streams and abort signals across child runs

This also links existing versioning and deploymentId: "latest" mentions back to the new Foundations page, and calls out that deploymentId is currently a Vercelism that may become a more generic version concept in the World spec later.

Follow-up cookbook pass added targeted callouts for long-lived or multi-turn patterns:

• scheduled workflows
• DurableAgent
• AI SDK multi-turn sessions
• Chat SDK sessions
• Sandbox sessions

The docs now also distinguish v5 from v4 for child-run starts. Native start() from workflow context was added by e295bae41 / #1491, which is on main but not stable, so v5 examples encourage direct workflow-context start() and v4 keeps the step-wrapped pattern.

Validation

• git diff --check
• Parsed the updated v4/v5 Foundations meta.json files as JSON
• Verified v4/v5 copies of the touched cookbook pages match before the v5-only native-start pass
• Searched docs/content for stale v5 step-wrapped start() guidance and v4 direct-start guidance
• Confirmed origin/stable lacks the 'use step' directive in packages/core/src/runtime/start.ts, while main has it via e295bae41 / feat: allow start() to be called directly inside workflow functions #1491

Full docs lint/typecheck were not run because this checkout does not have node_modules installed, so biome, vitest, and changeset are unavailable locally.

[vercel]

The latest updates on your projects. Learn more about Vercel for GitHub.

Project	Deployment	Actions	Updated (UTC)
example-nextjs-workflow-turbopack	Ready	Preview, Comment	May 18, 2026 10:49pm
example-nextjs-workflow-webpack	Ready	Preview, Comment	May 18, 2026 10:49pm
example-workflow	Ready	Preview, Comment	May 18, 2026 10:49pm
workbench-astro-workflow	Ready	Preview, Comment	May 18, 2026 10:49pm
workbench-express-workflow	Ready	Preview, Comment	May 18, 2026 10:49pm
workbench-fastify-workflow	Ready	Preview, Comment	May 18, 2026 10:49pm
workbench-hono-workflow	Ready	Preview, Comment	May 18, 2026 10:49pm
workbench-nitro-workflow	Ready	Preview, Comment	May 18, 2026 10:49pm
workbench-nuxt-workflow	Ready	Preview, Comment	May 18, 2026 10:49pm
workbench-sveltekit-workflow	Ready	Preview, Comment	May 18, 2026 10:49pm
workbench-tanstack-start-workflow	Ready	Preview, Comment	May 18, 2026 10:49pm
workbench-vite-workflow	Ready	Preview, Comment	May 18, 2026 10:49pm
workflow-docs	Ready	Preview, ✅ 11 resolved, Open in v0	May 18, 2026 10:49pm
workflow-swc-playground	Ready	Preview, Comment	May 18, 2026 10:49pm
workflow-tarballs	Ready	Preview, Comment	May 18, 2026 10:49pm
workflow-web	Ready	Preview, Comment	May 18, 2026 10:49pm

[changeset-bot]

🦋 Changeset detected

Latest commit: 7eff96c

The changes in this PR will be included in the next version bump.

This PR includes changesets to release 0 packages

When changesets are added to this PR, you'll see the packages that this PR includes changesets for and the associated semver types

Not sure what this means? Click here to learn what changesets are.

Click here if you're a maintainer who wants to add another changeset to this PR

[github-actions]

📊 Benchmark Results

📈 Comparing against baseline from main branch. Green 🟢 = faster, Red 🔺 = slower.

workflow with no steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
💻 Local	🥇 Nitro	0.031s (-28.3% 🟢)	1.005s (~)	0.974s	10	1.00x
💻 Local	Express	0.036s (-19.6% 🟢)	1.005s (~)	0.970s	10	1.15x
🐘 Postgres	Express	0.045s (-21.9% 🟢)	1.011s (~)	0.966s	10	1.47x
🐘 Postgres	Nitro	0.046s (-51.3% 🟢)	1.013s (-2.9%)	0.967s	10	1.50x

workflow with 1 step

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
💻 Local	🥇 Express	1.070s (-4.9%)	2.006s (~)	0.936s	10	1.00x
💻 Local	Nitro	1.071s (-5.3% 🟢)	2.006s (~)	0.935s	10	1.00x
🐘 Postgres	Express	1.079s (-5.9% 🟢)	2.009s (~)	0.930s	10	1.01x
🐘 Postgres	Nitro	1.081s (-5.2% 🟢)	2.010s (~)	0.929s	10	1.01x

workflow with 10 sequential steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Nitro	10.396s (-4.4%)	11.018s (~)	0.622s	3	1.00x
🐘 Postgres	Express	10.402s (-5.1% 🟢)	11.016s (~)	0.614s	3	1.00x
💻 Local	Express	10.411s (-4.7%)	11.023s (~)	0.612s	3	1.00x
💻 Local	Nitro	10.412s (-4.9%)	11.022s (~)	0.610s	3	1.00x

workflow with 25 sequential steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
💻 Local	🥇 Express	13.432s (-10.3% 🟢)	14.027s (-6.7% 🟢)	0.595s	5	1.00x
🐘 Postgres	Express	13.435s (-7.9% 🟢)	14.017s (-6.7% 🟢)	0.581s	5	1.00x
🐘 Postgres	Nitro	13.484s (-7.6% 🟢)	14.017s (-6.7% 🟢)	0.533s	5	1.00x
💻 Local	Nitro	13.489s (-10.4% 🟢)	14.027s (-12.5% 🟢)	0.538s	5	1.00x

workflow with 50 sequential steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
💻 Local	🥇 Nitro	11.875s (-29.2% 🟢)	12.022s (-29.4% 🟢)	0.147s	8	1.00x
💻 Local	Express	11.890s (-28.4% 🟢)	12.022s (-29.4% 🟢)	0.132s	8	1.00x
🐘 Postgres	Express	11.941s (-14.7% 🟢)	12.391s (-15.1% 🟢)	0.450s	8	1.01x
🐘 Postgres	Nitro	11.944s (-14.5% 🟢)	12.266s (-14.3% 🟢)	0.321s	8	1.01x

Promise.all with 10 concurrent steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Express	1.136s (-9.8% 🟢)	2.007s (~)	0.870s	15	1.00x
🐘 Postgres	Nitro	1.137s (-10.8% 🟢)	2.007s (~)	0.870s	15	1.00x
💻 Local	Nitro	1.172s (-28.2% 🟢)	2.006s (-3.3%)	0.834s	15	1.03x
💻 Local	Express	1.183s (-20.6% 🟢)	2.006s (~)	0.823s	15	1.04x

Promise.all with 25 concurrent steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Nitro	1.192s (-49.3% 🟢)	2.007s (-33.3% 🟢)	0.815s	15	1.00x
🐘 Postgres	Express	1.224s (-48.2% 🟢)	2.007s (-33.3% 🟢)	0.783s	15	1.03x
💻 Local	Express	1.703s (-42.3% 🟢)	2.005s (-41.9% 🟢)	0.303s	15	1.43x
💻 Local	Nitro	1.704s (-45.8% 🟢)	2.005s (-48.4% 🟢)	0.302s	15	1.43x

Promise.all with 50 concurrent steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Nitro	1.305s (-62.5% 🟢)	2.008s (-49.9% 🟢)	0.702s	15	1.00x
🐘 Postgres	Express	1.324s (-62.0% 🟢)	2.008s (-49.9% 🟢)	0.684s	15	1.01x
💻 Local	Express	4.638s (-44.4% 🟢)	5.179s (-42.6% 🟢)	0.541s	6	3.55x
💻 Local	Nitro	4.910s (-41.2% 🟢)	5.511s (-38.9% 🟢)	0.601s	6	3.76x

Promise.race with 10 concurrent steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Express	1.155s (-8.1% 🟢)	2.009s (~)	0.854s	15	1.00x
🐘 Postgres	Nitro	1.163s (-7.5% 🟢)	2.010s (~)	0.847s	15	1.01x
💻 Local	Express	1.386s (-26.8% 🟢)	2.006s (-15.2% 🟢)	0.620s	15	1.20x
💻 Local	Nitro	1.388s (-25.6% 🟢)	2.006s (-14.3% 🟢)	0.618s	15	1.20x

Promise.race with 25 concurrent steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Express	1.199s (-48.8% 🟢)	2.008s (-33.3% 🟢)	0.810s	15	1.00x
🐘 Postgres	Nitro	1.199s (-48.7% 🟢)	2.007s (-33.3% 🟢)	0.808s	15	1.00x
💻 Local	Express	1.828s (-41.6% 🟢)	2.222s (-40.9% 🟢)	0.394s	14	1.53x
💻 Local	Nitro	1.935s (-36.9% 🟢)	2.316s (-40.4% 🟢)	0.381s	13	1.61x

Promise.race with 50 concurrent steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Express	1.305s (-62.7% 🟢)	2.008s (-49.9% 🟢)	0.702s	15	1.00x
🐘 Postgres	Nitro	1.309s (-62.4% 🟢)	2.009s (-49.9% 🟢)	0.700s	15	1.00x
💻 Local	Express	5.446s (-38.1% 🟢)	5.849s (-36.9% 🟢)	0.403s	6	4.17x
💻 Local	Nitro	5.557s (-39.2% 🟢)	6.179s (-38.4% 🟢)	0.622s	6	4.26x

workflow with 10 sequential data payload steps (10KB)

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Express	0.418s (-50.1% 🟢)	1.006s (-1.6%)	0.588s	60	1.00x
🐘 Postgres	Nitro	0.459s (-44.0% 🟢)	1.006s (~)	0.547s	60	1.10x
💻 Local	Nitro	0.468s (-52.2% 🟢)	1.004s (-8.2% 🟢)	0.535s	60	1.12x
💻 Local	Express	0.469s (-52.4% 🟢)	1.004s (-6.7% 🟢)	0.535s	60	1.12x

workflow with 25 sequential data payload steps (10KB)

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Express	1.047s (-47.0% 🟢)	1.705s (-24.5% 🟢)	0.659s	53	1.00x
🐘 Postgres	Nitro	1.066s (-44.7% 🟢)	1.576s (-25.0% 🟢)	0.510s	58	1.02x
💻 Local	Nitro	1.185s (-61.0% 🟢)	2.006s (-46.6% 🟢)	0.821s	45	1.13x
💻 Local	Express	1.191s (-60.5% 🟢)	2.005s (-44.1% 🟢)	0.815s	45	1.14x

workflow with 50 sequential data payload steps (10KB)

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Express	2.030s (-49.1% 🟢)	2.427s (-44.4% 🟢)	0.397s	50	1.00x
🐘 Postgres	Nitro	2.128s (-48.2% 🟢)	2.660s (-42.2% 🟢)	0.532s	46	1.05x
💻 Local	Nitro	2.704s (-70.9% 🟢)	3.032s (-69.7% 🟢)	0.328s	40	1.33x
💻 Local	Express	2.719s (-70.5% 🟢)	3.057s (-69.5% 🟢)	0.338s	40	1.34x

workflow with 10 concurrent data payload steps (10KB)

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Express	0.163s (-42.4% 🟢)	1.005s (~)	0.843s	60	1.00x
🐘 Postgres	Nitro	0.178s (-37.0% 🟢)	1.006s (~)	0.827s	60	1.10x
💻 Local	Express	0.455s (-18.8% 🟢)	1.004s (~)	0.550s	60	2.80x
💻 Local	Nitro	0.469s (-22.4% 🟢)	1.004s (-1.7%)	0.535s	60	2.89x

workflow with 25 concurrent data payload steps (10KB)

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Express	0.286s (-43.9% 🟢)	1.006s (~)	0.720s	90	1.00x
🐘 Postgres	Nitro	0.295s (-40.5% 🟢)	1.006s (~)	0.711s	90	1.03x
💻 Local	Nitro	2.205s (-13.1% 🟢)	2.944s (-2.2%)	0.738s	31	7.71x
💻 Local	Express	2.226s (-11.4% 🟢)	2.943s (-2.2%)	0.717s	31	7.78x

workflow with 50 concurrent data payload steps (10KB)

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Express	0.582s (-28.9% 🟢)	1.007s (-1.0%)	0.425s	120	1.00x
🐘 Postgres	Nitro	0.634s (-19.7% 🟢)	1.006s (~)	0.372s	120	1.09x
💻 Local	Nitro	10.021s (-10.5% 🟢)	10.611s (-9.0% 🟢)	0.590s	12	17.21x
💻 Local	Express	10.042s (-10.3% 🟢)	10.694s (-10.4% 🟢)	0.652s	12	17.25x

Stream Benchmarks(includes TTFB metrics)

workflow with stream

💻 Local Development

World	Framework	Workflow Time	TTFB	Slurp	Wall Time	Overhead	Samples	vs Fastest
💻 Local	🥇 Express	1.132s (+468.5% 🔺)	2.005s (+99.6% 🔺)	0.011s (-13.2% 🟢)	2.018s (+98.2% 🔺)	0.886s	10	1.00x
💻 Local	Nitro	1.133s (+430.4% 🔺)	2.005s (+99.6% 🔺)	0.011s (-15.2% 🟢)	2.018s (+98.1% 🔺)	0.885s	10	1.00x
🐘 Postgres	Nitro	1.135s (+453.6% 🔺)	2.001s (+100.2% 🔺)	0.001s (-20.0% 🟢)	2.010s (+98.7% 🔺)	0.875s	10	1.00x
🐘 Postgres	Express	1.135s (+453.5% 🔺)	1.999s (+100.2% 🔺)	0.001s (-25.0% 🟢)	2.010s (+98.7% 🔺)	0.875s	10	1.00x

stream pipeline with 5 transform steps (1MB)

💻 Local Development

World	Framework	Workflow Time	TTFB	Slurp	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Express	1.492s (+136.8% 🔺)	2.002s (+98.9% 🔺)	0.003s (-30.4% 🟢)	2.023s (+97.8% 🔺)	0.531s	30	1.00x
🐘 Postgres	Nitro	1.519s (+143.4% 🔺)	2.004s (+99.0% 🔺)	0.004s (-2.5%)	2.027s (+98.2% 🔺)	0.507s	30	1.02x
💻 Local	Nitro	1.707s (+103.5% 🔺)	2.013s (+98.9% 🔺)	0.009s (-4.3%)	2.203s (+97.4% 🔺)	0.496s	28	1.14x
💻 Local	Express	1.923s (+154.0% 🔺)	2.011s (+95.4% 🔺)	0.009s (-8.2% 🟢)	2.422s (+132.9% 🔺)	0.499s	25	1.29x

10 parallel streams (1MB each)

💻 Local Development

World	Framework	Workflow Time	TTFB	Slurp	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Express	0.627s (-34.7% 🟢)	1.033s (-19.2% 🟢)	0.000s (-20.7% 🟢)	1.045s (-20.0% 🟢)	0.418s	58	1.00x
🐘 Postgres	Nitro	0.662s (-31.7% 🟢)	1.029s (-17.5% 🟢)	0.000s (-17.2% 🟢)	1.049s (-16.6% 🟢)	0.387s	58	1.05x
💻 Local	Express	1.365s (+11.5% 🔺)	2.015s (~)	0.000s (+40.0% 🔺)	2.018s (~)	0.652s	30	2.18x
💻 Local	Nitro	1.561s (+27.6% 🔺)	2.016s (~)	0.000s (+364.3% 🔺)	2.197s (+8.7% 🔺)	0.636s	28	2.49x

fan-out fan-in 10 streams (1MB each)

💻 Local Development

World	Framework	Workflow Time	TTFB	Slurp	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Express	1.228s (-30.7% 🟢)	1.869s (-14.2% 🟢)	0.000s (NaN%)	1.900s (-13.6% 🟢)	0.673s	32	1.00x
🐘 Postgres	Nitro	1.325s (-26.1% 🟢)	1.997s (-6.7% 🟢)	0.000s (+86.7% 🔺)	2.019s (-7.2% 🟢)	0.694s	30	1.08x
💻 Local	Express	3.158s (-8.9% 🟢)	4.028s (~)	0.001s (-33.3% 🟢)	4.030s (~)	0.873s	15	2.57x
💻 Local	Nitro	3.187s (-5.9% 🟢)	4.029s (~)	0.000s (-75.0% 🟢)	4.031s (~)	0.844s	15	2.60x

Summary

Fastest Framework by World

Winner determined by most benchmark wins

World	🥇 Fastest Framework	Wins
💻 Local	Express	12/21
🐘 Postgres	Express	17/21

Fastest World by Framework

Winner determined by most benchmark wins

Framework	🥇 Fastest World	Wins
Express	🐘 Postgres	16/21
Nitro	🐘 Postgres	17/21

Column Definitions

• Workflow Time: Runtime reported by workflow (completedAt - createdAt) - primary metric
• TTFB: Time to First Byte - time from workflow start until first stream byte received (stream benchmarks only)
• Slurp: Time from first byte to complete stream consumption (stream benchmarks only)
• Wall Time: Total testbench time (trigger workflow + poll for result)
• Overhead: Testbench overhead (Wall Time - Workflow Time)
• Samples: Number of benchmark iterations run
• vs Fastest: How much slower compared to the fastest configuration for this benchmark

Worlds:

• 💻 Local: In-memory filesystem world (local development)
• 🐘 Postgres: PostgreSQL database world (local development)
• ▲ Vercel: Vercel production/preview deployment
• 🌐 Turso: Community world (local development)
• 🌐 MongoDB: Community world (local development)
• 🌐 Redis: Community world (local development)
• 🌐 Jazz: Community world (local development)
• 🌐 Redis: Community world (local development)
• 🌐 Redis + BullMQ: Community world (local development)
• 🌐 Cloudflare: Community world (local development)
• 🌐 MySQL: Community world (local development)
• 🌐 Azure: Community world (local development)
• 🌐 NATS JetStream: Community world (local development)
• 🌐 Upstash: Community world (local development)

---

📋 View full workflow run

---

❌ Some benchmark jobs failed:

• Local: success
• Postgres: success
• Vercel: failure

Check the workflow run for details.

[github-actions]

🧪 E2E Test Results

❌ Some tests failed

Summary

	Passed	Failed	Skipped	Total
❌ ▲ Vercel Production	1196	4	219	1419
✅ 💻 Local Development	1587	0	219	1806
✅ 📦 Local Production	1458	0	219	1677
❌ 🐘 Local Postgres	1585	2	219	1806
✅ 🪟 Windows	129	0	0	129
✅ 📋 Other	727	0	176	903
Total	6682	6	1052	7740

❌ Failed Tests

▲ Vercel Production (4 failed)

astro (1 failed):

• AbortController abortFromStepWorkflow: step abort cancels an in-flight sibling step

hono (2 failed):

• outputStreamWorkflow no startIndex (reads all chunks)
• stepFunctionPassingWorkflow - step function references can be passed as arguments (without closure vars) | wrun_01KRYMPWTNKF0CE4BP27X7R946 | 🔍 observability

nextjs-webpack (1 failed):

• parallelSleepWorkflow | wrun_01KRYME5H0F7D8EH2AAF56T0Z2 | 🔍 observability

🐘 Local Postgres (2 failed)

nextjs-webpack-stable-lazy-discovery-disabled (1 failed):

• DurableAgent e2e core basic text response

nextjs-webpack-stable-lazy-discovery-enabled (1 failed):

• addTenWorkflow | wrun_01KRYMBT8QHFK9EZ9FDY03D9G0

Details by Category

❌ ▲ Vercel Production

App	Passed	Failed	Skipped
❌ astro	102	1	26
✅ example	103	0	26
✅ express	103	0	26
✅ fastify	103	0	26
❌ hono	101	2	26
✅ nextjs-turbopack	127	0	2
❌ nextjs-webpack	126	1	2
✅ nitro	103	0	26
✅ nuxt	103	0	26
✅ sveltekit	122	0	7
✅ vite	103	0	26

✅ 💻 Local Development

App	Passed	Failed	Skipped
✅ astro-stable	104	0	25
✅ express-stable	104	0	25
✅ fastify-stable	104	0	25
✅ hono-stable	104	0	25
✅ nextjs-turbopack-canary	110	0	19
✅ nextjs-turbopack-stable-lazy-discovery-disabled	129	0	0
✅ nextjs-turbopack-stable-lazy-discovery-enabled	129	0	0
✅ nextjs-webpack-canary	110	0	19
✅ nextjs-webpack-stable-lazy-discovery-disabled	129	0	0
✅ nextjs-webpack-stable-lazy-discovery-enabled	129	0	0
✅ nitro-stable	104	0	25
✅ nuxt-stable	104	0	25
✅ sveltekit-stable	123	0	6
✅ vite-stable	104	0	25

✅ 📦 Local Production

App	Passed	Failed	Skipped
✅ astro-stable	104	0	25
✅ express-stable	104	0	25
✅ fastify-stable	104	0	25
✅ hono-stable	104	0	25
✅ nextjs-turbopack-canary	110	0	19
✅ nextjs-turbopack-stable-lazy-discovery-disabled	129	0	0
✅ nextjs-webpack-canary	110	0	19
✅ nextjs-webpack-stable-lazy-discovery-disabled	129	0	0
✅ nextjs-webpack-stable-lazy-discovery-enabled	129	0	0
✅ nitro-stable	104	0	25
✅ nuxt-stable	104	0	25
✅ sveltekit-stable	123	0	6
✅ vite-stable	104	0	25

❌ 🐘 Local Postgres

App	Passed	Failed	Skipped
✅ astro-stable	104	0	25
✅ express-stable	104	0	25
✅ fastify-stable	104	0	25
✅ hono-stable	104	0	25
✅ nextjs-turbopack-canary	110	0	19
✅ nextjs-turbopack-stable-lazy-discovery-disabled	129	0	0
✅ nextjs-turbopack-stable-lazy-discovery-enabled	129	0	0
✅ nextjs-webpack-canary	110	0	19
❌ nextjs-webpack-stable-lazy-discovery-disabled	128	1	0
❌ nextjs-webpack-stable-lazy-discovery-enabled	128	1	0
✅ nitro-stable	104	0	25
✅ nuxt-stable	104	0	25
✅ sveltekit-stable	123	0	6
✅ vite-stable	104	0	25

✅ 🪟 Windows

App	Passed	Failed	Skipped
✅ nextjs-turbopack	129	0	0

✅ 📋 Other

App	Passed	Failed	Skipped
✅ e2e-local-dev-nest-stable	104	0	25
✅ e2e-local-dev-tanstack-start-	104	0	25
✅ e2e-local-postgres-nest-stable	104	0	25
✅ e2e-local-postgres-tanstack-start-	104	0	25
✅ e2e-local-prod-nest-stable	104	0	25
✅ e2e-local-prod-tanstack-start-	104	0	25
✅ e2e-vercel-prod-tanstack-start	103	0	26

---

📋 View full workflow run

---

❌ Some E2E test jobs failed:

• Vercel Prod: failure
• Local Dev: success
• Local Prod: failure
• Local Postgres: failure
• Windows: success

Check the workflow run for details.

[Copilot]

Pull request overview

Adds a new "Versioning" Foundations page (v4 and v5) explaining default run pinning to deployments, recovery via cancel + rerun, opting into newer code with deploymentId: "latest", and carrying serializable state/streams across child-run continuations. Existing references to deploymentId: "latest" and versioning behavior are linked back to the new page. Cookbook pages also pick up callouts about long-lived / multi-turn patterns and, on v5 only, are updated to reflect that start() can now be called directly inside "use workflow" functions (via #1491), while v4 retains the step-wrapped pattern.

Changes:

• New foundations/versioning.mdx for both v4 and v5, plus meta.json + index.mdx + linkbacks from vercel-world, start.mdx, and code-transform.mdx.
• v5 cookbook + migration guides updated to drop "use step" wrappers around start() in favor of calling start() directly from the workflow function; v4 keeps the step-wrapped pattern.
• Added "Versioning" callouts to scheduling, durable-agent, ai-sdk, chat-sdk, and sandbox cookbooks; fixed a few small wording/grammar issues in code-transform.mdx.

Reviewed changes

Copilot reviewed 32 out of 32 changed files in this pull request and generated no comments.

Show a summary per file

File	Description
docs/content/docs/v5/foundations/versioning.mdx	New Foundations page explaining run pinning, recovery, and explicit upgrade boundaries.
docs/content/docs/v4/foundations/versioning.mdx	v4 counterpart; uses step-wrapped continueDigest/continueExportOnLatest helpers.
docs/content/docs/v{4,5}/foundations/{meta.json,index.mdx}	Surface the new Versioning page in nav and the foundations index.
docs/content/docs/v{4,5}/deploying/world/vercel-world.mdx	Link the deployment-pinning section to the new Versioning page.
docs/content/docs/v{4,5}/how-it-works/code-transform.mdx	Link "atomic versioning" reference; fix "function assume" / "from being upgraded" wording.
docs/content/docs/v{4,5}/api-reference/workflow-api/start.mdx	Soften "Vercel-specific" to "Vercelism", mention possible version rename, link Versioning. v5 also notes native workflow-context start().
docs/content/docs/v5/foundations/starting-workflows.mdx	Note v5 can call start() from workflows; link Composition + Versioning.
docs/content/docs/v5/cookbook/common-patterns/workflow-composition.mdx	Replace step-wrapped child spawn with direct start() call.
docs/content/docs/v5/cookbook/advanced/child-workflows.mdx	Inline batched start() calls; rename spawnReportChunk → startReportChunk; update tips.
docs/content/docs/v{4,5}/cookbook/common-patterns/scheduling.mdx	Callout linking to Versioning for long-lived schedules.
docs/content/docs/v{4,5}/cookbook/agent-patterns/durable-agent.mdx	Callout linking to Versioning for multi-turn agents.
docs/content/docs/v{4,5}/cookbook/integrations/{ai-sdk,chat-sdk,sandbox}.mdx	Callouts linking to Versioning for long-lived sessions.
docs/content/docs/v4/cookbook/common-patterns/workflow-composition.mdx	Update deploymentId: "latest" callout wording + link to Versioning.
docs/content/docs/v4/cookbook/advanced/child-workflows.mdx	Link Versioning from the deploymentId: "latest" tip.
docs/content/docs/v4/changelog/eager-processing.mdx	Reword to "step-wrapped start() helper" to match v4 phrasing.
docs/content/docs/v5/migration-guides/{migrating-from-trigger-dev,migrating-from-temporal,migrating-from-inngest,migrating-from-aws-step-functions}.mdx	Update to v5 native workflow-context start(); only collect step retains "use step".

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[TooTallNate]

Comment-only review

The new versioning foundations page is a useful addition — the four sections (default pinning → recovery → self-upgrade via start({ deploymentId: "latest" }) → carrying serializable context forward) cover the model clearly, and the v4-vs-v5 differentiation (step-wrapped start() vs direct start() from workflow context, gated on whether PR #1491 has landed on the target branch) is correctly handled. The cross-links from cookbook patterns to the new foundations page tie everything together nicely.

Three observations — see inline.

Inline comments

1. Wording: this PR introduces "Vercelism" as a new term and uses it 6 times across the diff, including replacing existing "Vercel-specific feature" wording in two <Callout>s. "Vercelism" reads as informal slang and isn't part of the established docs vocabulary. The existing tone uses "Vercel-specific", "Vercel deployments", "Vercel World", "non-Vercel" — all precise. Suggested restoring the prior phrasing. This is the only thing I'd consider blocking.
2. Broken link in v4: versioning.mdx references /docs/foundations/cancellation in its related frontmatter, but that page is v5-only. Easy fix.
3. Missing changeset: convention for docs-only PRs is an empty ---\n--- file.

A few small things I noticed

• The pre-emptively check for a conflict with getHookByToken() example (line 96 of versioning.mdx) — nice that you noted the race window in a later doc (#2011 callout), but this page doesn't have that warning. Worth a cross-link or <Callout type="warn"> here too, since the recovery pattern uses both start() and deploymentId: "latest" together.
• The _idempotency / using syntax is great. Worth confirming the docs site's syntax highlighter renders it — I noticed #1988 recently fixed a homepage link lint regression, and the using syntax is TypeScript 5.2+ specific.
• dailyDigest example (line 137-152): the state parameter is DigestState but the workflow doesn't actually use state.userId after the first sendDigest call — pattern is fine, just noting if you wanted to make the example illustrate state evolution more vividly, having sendDigest read state.lastSentAt (e.g. to avoid sending duplicates) would make the carry-forward intent clearer.

Overall the page reads well — once the Vercelism wording is restored and the v4 link is fixed, this should be good to go.

[TooTallNate]

All three review comments addressed in 7eff96c — Vercelism wording reverted, v4 broken cancellation link fixed, empty changeset added. LGTM, approving.

[github-actions]

Backport PR opened against stable: #2014. Merge conflicts were resolved by AI — please review carefully.

[pranaygp]

Summary

Adds a new Foundations page explaining Workflow SDK versioning semantics for both v5 and v4 docs.

The page covers:

• default run pinning to the deployment that started the run
• canceling and rerunning affected in-flight runs on the latest deployment
• explicit self-upgrade patterns using deploymentId: "latest"
• carrying serializable context such as streams and abort signals across child runs

This also links existing versioning and deploymentId: "latest" mentions back to the new Foundations page, and calls out that deploymentId is currently a Vercelism that may become a more generic version concept in the World spec later.

Follow-up cookbook pass added targeted callouts for long-lived or multi-turn patterns:

• scheduled workflows
• DurableAgent
• AI SDK multi-turn sessions
• Chat SDK sessions
• Sandbox sessions

The docs now also distinguish v5 from v4 for child-run starts. Native start() from workflow context was added by e295bae41 / #1491, which is on main but not stable, so v5 examples encourage direct workflow-context start() and v4 keeps the step-wrapped pattern.

Validation

• git diff --check
• Parsed the updated v4/v5 Foundations meta.json files as JSON
• Verified v4/v5 copies of the touched cookbook pages match before the v5-only native-start pass
• Searched docs/content for stale v5 step-wrapped start() guidance and v4 direct-start guidance
• Confirmed origin/stable lacks the 'use step' directive in packages/core/src/runtime/start.ts, while main has it via e295bae41 / feat: allow start() to be called directly inside workflow functions #1491

Full docs lint/typecheck were not run because this checkout does not have node_modules installed, so biome, vitest, and changeset are unavailable locally.

[vercel]

The latest updates on your projects. Learn more about Vercel for GitHub.

Project	Deployment	Actions	Updated (UTC)
example-nextjs-workflow-turbopack	Ready	Preview, Comment	May 18, 2026 10:49pm
example-nextjs-workflow-webpack	Ready	Preview, Comment	May 18, 2026 10:49pm
example-workflow	Ready	Preview, Comment	May 18, 2026 10:49pm
workbench-astro-workflow	Ready	Preview, Comment	May 18, 2026 10:49pm
workbench-express-workflow	Ready	Preview, Comment	May 18, 2026 10:49pm
workbench-fastify-workflow	Ready	Preview, Comment	May 18, 2026 10:49pm
workbench-hono-workflow	Ready	Preview, Comment	May 18, 2026 10:49pm
workbench-nitro-workflow	Ready	Preview, Comment	May 18, 2026 10:49pm
workbench-nuxt-workflow	Ready	Preview, Comment	May 18, 2026 10:49pm
workbench-sveltekit-workflow	Ready	Preview, Comment	May 18, 2026 10:49pm
workbench-tanstack-start-workflow	Ready	Preview, Comment	May 18, 2026 10:49pm
workbench-vite-workflow	Ready	Preview, Comment	May 18, 2026 10:49pm
workflow-docs	Ready	Preview, ✅ 11 resolved, Open in v0	May 18, 2026 10:49pm
workflow-swc-playground	Ready	Preview, Comment	May 18, 2026 10:49pm
workflow-tarballs	Ready	Preview, Comment	May 18, 2026 10:49pm
workflow-web	Ready	Preview, Comment	May 18, 2026 10:49pm

[changeset-bot]

🦋 Changeset detected

Latest commit: 7eff96c

The changes in this PR will be included in the next version bump.

This PR includes changesets to release 0 packages

When changesets are added to this PR, you'll see the packages that this PR includes changesets for and the associated semver types

Not sure what this means? Click here to learn what changesets are.

Click here if you're a maintainer who wants to add another changeset to this PR

[github-actions]

📊 Benchmark Results

📈 Comparing against baseline from main branch. Green 🟢 = faster, Red 🔺 = slower.

workflow with no steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
💻 Local	🥇 Nitro	0.031s (-28.3% 🟢)	1.005s (~)	0.974s	10	1.00x
💻 Local	Express	0.036s (-19.6% 🟢)	1.005s (~)	0.970s	10	1.15x
🐘 Postgres	Express	0.045s (-21.9% 🟢)	1.011s (~)	0.966s	10	1.47x
🐘 Postgres	Nitro	0.046s (-51.3% 🟢)	1.013s (-2.9%)	0.967s	10	1.50x

workflow with 1 step

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
💻 Local	🥇 Express	1.070s (-4.9%)	2.006s (~)	0.936s	10	1.00x
💻 Local	Nitro	1.071s (-5.3% 🟢)	2.006s (~)	0.935s	10	1.00x
🐘 Postgres	Express	1.079s (-5.9% 🟢)	2.009s (~)	0.930s	10	1.01x
🐘 Postgres	Nitro	1.081s (-5.2% 🟢)	2.010s (~)	0.929s	10	1.01x

workflow with 10 sequential steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Nitro	10.396s (-4.4%)	11.018s (~)	0.622s	3	1.00x
🐘 Postgres	Express	10.402s (-5.1% 🟢)	11.016s (~)	0.614s	3	1.00x
💻 Local	Express	10.411s (-4.7%)	11.023s (~)	0.612s	3	1.00x
💻 Local	Nitro	10.412s (-4.9%)	11.022s (~)	0.610s	3	1.00x

workflow with 25 sequential steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
💻 Local	🥇 Express	13.432s (-10.3% 🟢)	14.027s (-6.7% 🟢)	0.595s	5	1.00x
🐘 Postgres	Express	13.435s (-7.9% 🟢)	14.017s (-6.7% 🟢)	0.581s	5	1.00x
🐘 Postgres	Nitro	13.484s (-7.6% 🟢)	14.017s (-6.7% 🟢)	0.533s	5	1.00x
💻 Local	Nitro	13.489s (-10.4% 🟢)	14.027s (-12.5% 🟢)	0.538s	5	1.00x

workflow with 50 sequential steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
💻 Local	🥇 Nitro	11.875s (-29.2% 🟢)	12.022s (-29.4% 🟢)	0.147s	8	1.00x
💻 Local	Express	11.890s (-28.4% 🟢)	12.022s (-29.4% 🟢)	0.132s	8	1.00x
🐘 Postgres	Express	11.941s (-14.7% 🟢)	12.391s (-15.1% 🟢)	0.450s	8	1.01x
🐘 Postgres	Nitro	11.944s (-14.5% 🟢)	12.266s (-14.3% 🟢)	0.321s	8	1.01x

Promise.all with 10 concurrent steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Express	1.136s (-9.8% 🟢)	2.007s (~)	0.870s	15	1.00x
🐘 Postgres	Nitro	1.137s (-10.8% 🟢)	2.007s (~)	0.870s	15	1.00x
💻 Local	Nitro	1.172s (-28.2% 🟢)	2.006s (-3.3%)	0.834s	15	1.03x
💻 Local	Express	1.183s (-20.6% 🟢)	2.006s (~)	0.823s	15	1.04x

Promise.all with 25 concurrent steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Nitro	1.192s (-49.3% 🟢)	2.007s (-33.3% 🟢)	0.815s	15	1.00x
🐘 Postgres	Express	1.224s (-48.2% 🟢)	2.007s (-33.3% 🟢)	0.783s	15	1.03x
💻 Local	Express	1.703s (-42.3% 🟢)	2.005s (-41.9% 🟢)	0.303s	15	1.43x
💻 Local	Nitro	1.704s (-45.8% 🟢)	2.005s (-48.4% 🟢)	0.302s	15	1.43x

Promise.all with 50 concurrent steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Nitro	1.305s (-62.5% 🟢)	2.008s (-49.9% 🟢)	0.702s	15	1.00x
🐘 Postgres	Express	1.324s (-62.0% 🟢)	2.008s (-49.9% 🟢)	0.684s	15	1.01x
💻 Local	Express	4.638s (-44.4% 🟢)	5.179s (-42.6% 🟢)	0.541s	6	3.55x
💻 Local	Nitro	4.910s (-41.2% 🟢)	5.511s (-38.9% 🟢)	0.601s	6	3.76x

Promise.race with 10 concurrent steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Express	1.155s (-8.1% 🟢)	2.009s (~)	0.854s	15	1.00x
🐘 Postgres	Nitro	1.163s (-7.5% 🟢)	2.010s (~)	0.847s	15	1.01x
💻 Local	Express	1.386s (-26.8% 🟢)	2.006s (-15.2% 🟢)	0.620s	15	1.20x
💻 Local	Nitro	1.388s (-25.6% 🟢)	2.006s (-14.3% 🟢)	0.618s	15	1.20x

Promise.race with 25 concurrent steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Express	1.199s (-48.8% 🟢)	2.008s (-33.3% 🟢)	0.810s	15	1.00x
🐘 Postgres	Nitro	1.199s (-48.7% 🟢)	2.007s (-33.3% 🟢)	0.808s	15	1.00x
💻 Local	Express	1.828s (-41.6% 🟢)	2.222s (-40.9% 🟢)	0.394s	14	1.53x
💻 Local	Nitro	1.935s (-36.9% 🟢)	2.316s (-40.4% 🟢)	0.381s	13	1.61x

Promise.race with 50 concurrent steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Express	1.305s (-62.7% 🟢)	2.008s (-49.9% 🟢)	0.702s	15	1.00x
🐘 Postgres	Nitro	1.309s (-62.4% 🟢)	2.009s (-49.9% 🟢)	0.700s	15	1.00x
💻 Local	Express	5.446s (-38.1% 🟢)	5.849s (-36.9% 🟢)	0.403s	6	4.17x
💻 Local	Nitro	5.557s (-39.2% 🟢)	6.179s (-38.4% 🟢)	0.622s	6	4.26x

workflow with 10 sequential data payload steps (10KB)

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Express	0.418s (-50.1% 🟢)	1.006s (-1.6%)	0.588s	60	1.00x
🐘 Postgres	Nitro	0.459s (-44.0% 🟢)	1.006s (~)	0.547s	60	1.10x
💻 Local	Nitro	0.468s (-52.2% 🟢)	1.004s (-8.2% 🟢)	0.535s	60	1.12x
💻 Local	Express	0.469s (-52.4% 🟢)	1.004s (-6.7% 🟢)	0.535s	60	1.12x

workflow with 25 sequential data payload steps (10KB)

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Express	1.047s (-47.0% 🟢)	1.705s (-24.5% 🟢)	0.659s	53	1.00x
🐘 Postgres	Nitro	1.066s (-44.7% 🟢)	1.576s (-25.0% 🟢)	0.510s	58	1.02x
💻 Local	Nitro	1.185s (-61.0% 🟢)	2.006s (-46.6% 🟢)	0.821s	45	1.13x
💻 Local	Express	1.191s (-60.5% 🟢)	2.005s (-44.1% 🟢)	0.815s	45	1.14x

workflow with 50 sequential data payload steps (10KB)

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Express	2.030s (-49.1% 🟢)	2.427s (-44.4% 🟢)	0.397s	50	1.00x
🐘 Postgres	Nitro	2.128s (-48.2% 🟢)	2.660s (-42.2% 🟢)	0.532s	46	1.05x
💻 Local	Nitro	2.704s (-70.9% 🟢)	3.032s (-69.7% 🟢)	0.328s	40	1.33x
💻 Local	Express	2.719s (-70.5% 🟢)	3.057s (-69.5% 🟢)	0.338s	40	1.34x

workflow with 10 concurrent data payload steps (10KB)

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Express	0.163s (-42.4% 🟢)	1.005s (~)	0.843s	60	1.00x
🐘 Postgres	Nitro	0.178s (-37.0% 🟢)	1.006s (~)	0.827s	60	1.10x
💻 Local	Express	0.455s (-18.8% 🟢)	1.004s (~)	0.550s	60	2.80x
💻 Local	Nitro	0.469s (-22.4% 🟢)	1.004s (-1.7%)	0.535s	60	2.89x

workflow with 25 concurrent data payload steps (10KB)

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Express	0.286s (-43.9% 🟢)	1.006s (~)	0.720s	90	1.00x
🐘 Postgres	Nitro	0.295s (-40.5% 🟢)	1.006s (~)	0.711s	90	1.03x
💻 Local	Nitro	2.205s (-13.1% 🟢)	2.944s (-2.2%)	0.738s	31	7.71x
💻 Local	Express	2.226s (-11.4% 🟢)	2.943s (-2.2%)	0.717s	31	7.78x

workflow with 50 concurrent data payload steps (10KB)

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Express	0.582s (-28.9% 🟢)	1.007s (-1.0%)	0.425s	120	1.00x
🐘 Postgres	Nitro	0.634s (-19.7% 🟢)	1.006s (~)	0.372s	120	1.09x
💻 Local	Nitro	10.021s (-10.5% 🟢)	10.611s (-9.0% 🟢)	0.590s	12	17.21x
💻 Local	Express	10.042s (-10.3% 🟢)	10.694s (-10.4% 🟢)	0.652s	12	17.25x

Stream Benchmarks(includes TTFB metrics)

workflow with stream

💻 Local Development

World	Framework	Workflow Time	TTFB	Slurp	Wall Time	Overhead	Samples	vs Fastest
💻 Local	🥇 Express	1.132s (+468.5% 🔺)	2.005s (+99.6% 🔺)	0.011s (-13.2% 🟢)	2.018s (+98.2% 🔺)	0.886s	10	1.00x
💻 Local	Nitro	1.133s (+430.4% 🔺)	2.005s (+99.6% 🔺)	0.011s (-15.2% 🟢)	2.018s (+98.1% 🔺)	0.885s	10	1.00x
🐘 Postgres	Nitro	1.135s (+453.6% 🔺)	2.001s (+100.2% 🔺)	0.001s (-20.0% 🟢)	2.010s (+98.7% 🔺)	0.875s	10	1.00x
🐘 Postgres	Express	1.135s (+453.5% 🔺)	1.999s (+100.2% 🔺)	0.001s (-25.0% 🟢)	2.010s (+98.7% 🔺)	0.875s	10	1.00x

stream pipeline with 5 transform steps (1MB)

💻 Local Development

World	Framework	Workflow Time	TTFB	Slurp	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Express	1.492s (+136.8% 🔺)	2.002s (+98.9% 🔺)	0.003s (-30.4% 🟢)	2.023s (+97.8% 🔺)	0.531s	30	1.00x
🐘 Postgres	Nitro	1.519s (+143.4% 🔺)	2.004s (+99.0% 🔺)	0.004s (-2.5%)	2.027s (+98.2% 🔺)	0.507s	30	1.02x
💻 Local	Nitro	1.707s (+103.5% 🔺)	2.013s (+98.9% 🔺)	0.009s (-4.3%)	2.203s (+97.4% 🔺)	0.496s	28	1.14x
💻 Local	Express	1.923s (+154.0% 🔺)	2.011s (+95.4% 🔺)	0.009s (-8.2% 🟢)	2.422s (+132.9% 🔺)	0.499s	25	1.29x

10 parallel streams (1MB each)

💻 Local Development

World	Framework	Workflow Time	TTFB	Slurp	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Express	0.627s (-34.7% 🟢)	1.033s (-19.2% 🟢)	0.000s (-20.7% 🟢)	1.045s (-20.0% 🟢)	0.418s	58	1.00x
🐘 Postgres	Nitro	0.662s (-31.7% 🟢)	1.029s (-17.5% 🟢)	0.000s (-17.2% 🟢)	1.049s (-16.6% 🟢)	0.387s	58	1.05x
💻 Local	Express	1.365s (+11.5% 🔺)	2.015s (~)	0.000s (+40.0% 🔺)	2.018s (~)	0.652s	30	2.18x
💻 Local	Nitro	1.561s (+27.6% 🔺)	2.016s (~)	0.000s (+364.3% 🔺)	2.197s (+8.7% 🔺)	0.636s	28	2.49x

fan-out fan-in 10 streams (1MB each)

💻 Local Development

World	Framework	Workflow Time	TTFB	Slurp	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Express	1.228s (-30.7% 🟢)	1.869s (-14.2% 🟢)	0.000s (NaN%)	1.900s (-13.6% 🟢)	0.673s	32	1.00x
🐘 Postgres	Nitro	1.325s (-26.1% 🟢)	1.997s (-6.7% 🟢)	0.000s (+86.7% 🔺)	2.019s (-7.2% 🟢)	0.694s	30	1.08x
💻 Local	Express	3.158s (-8.9% 🟢)	4.028s (~)	0.001s (-33.3% 🟢)	4.030s (~)	0.873s	15	2.57x
💻 Local	Nitro	3.187s (-5.9% 🟢)	4.029s (~)	0.000s (-75.0% 🟢)	4.031s (~)	0.844s	15	2.60x

Summary

Fastest Framework by World

Winner determined by most benchmark wins

World	🥇 Fastest Framework	Wins
💻 Local	Express	12/21
🐘 Postgres	Express	17/21

Fastest World by Framework

Winner determined by most benchmark wins

Framework	🥇 Fastest World	Wins
Express	🐘 Postgres	16/21
Nitro	🐘 Postgres	17/21

Column Definitions

• Workflow Time: Runtime reported by workflow (completedAt - createdAt) - primary metric
• TTFB: Time to First Byte - time from workflow start until first stream byte received (stream benchmarks only)
• Slurp: Time from first byte to complete stream consumption (stream benchmarks only)
• Wall Time: Total testbench time (trigger workflow + poll for result)
• Overhead: Testbench overhead (Wall Time - Workflow Time)
• Samples: Number of benchmark iterations run
• vs Fastest: How much slower compared to the fastest configuration for this benchmark

Worlds:

• 💻 Local: In-memory filesystem world (local development)
• 🐘 Postgres: PostgreSQL database world (local development)
• ▲ Vercel: Vercel production/preview deployment
• 🌐 Turso: Community world (local development)
• 🌐 MongoDB: Community world (local development)
• 🌐 Redis: Community world (local development)
• 🌐 Jazz: Community world (local development)
• 🌐 Redis: Community world (local development)
• 🌐 Redis + BullMQ: Community world (local development)
• 🌐 Cloudflare: Community world (local development)
• 🌐 MySQL: Community world (local development)
• 🌐 Azure: Community world (local development)
• 🌐 NATS JetStream: Community world (local development)
• 🌐 Upstash: Community world (local development)

---

📋 View full workflow run

---

❌ Some benchmark jobs failed:

• Local: success
• Postgres: success
• Vercel: failure

Check the workflow run for details.

[github-actions]

🧪 E2E Test Results

❌ Some tests failed

Summary

	Passed	Failed	Skipped	Total
❌ ▲ Vercel Production	1196	4	219	1419
✅ 💻 Local Development	1587	0	219	1806
✅ 📦 Local Production	1458	0	219	1677
❌ 🐘 Local Postgres	1585	2	219	1806
✅ 🪟 Windows	129	0	0	129
✅ 📋 Other	727	0	176	903
Total	6682	6	1052	7740

❌ Failed Tests

▲ Vercel Production (4 failed)

astro (1 failed):

• AbortController abortFromStepWorkflow: step abort cancels an in-flight sibling step

hono (2 failed):

• outputStreamWorkflow no startIndex (reads all chunks)
• stepFunctionPassingWorkflow - step function references can be passed as arguments (without closure vars) | wrun_01KRYMPWTNKF0CE4BP27X7R946 | 🔍 observability

nextjs-webpack (1 failed):

• parallelSleepWorkflow | wrun_01KRYME5H0F7D8EH2AAF56T0Z2 | 🔍 observability

🐘 Local Postgres (2 failed)

nextjs-webpack-stable-lazy-discovery-disabled (1 failed):

• DurableAgent e2e core basic text response

nextjs-webpack-stable-lazy-discovery-enabled (1 failed):

• addTenWorkflow | wrun_01KRYMBT8QHFK9EZ9FDY03D9G0

Details by Category

❌ ▲ Vercel Production

App	Passed	Failed	Skipped
❌ astro	102	1	26
✅ example	103	0	26
✅ express	103	0	26
✅ fastify	103	0	26
❌ hono	101	2	26
✅ nextjs-turbopack	127	0	2
❌ nextjs-webpack	126	1	2
✅ nitro	103	0	26
✅ nuxt	103	0	26
✅ sveltekit	122	0	7
✅ vite	103	0	26

✅ 💻 Local Development

App	Passed	Failed	Skipped
✅ astro-stable	104	0	25
✅ express-stable	104	0	25
✅ fastify-stable	104	0	25
✅ hono-stable	104	0	25
✅ nextjs-turbopack-canary	110	0	19
✅ nextjs-turbopack-stable-lazy-discovery-disabled	129	0	0
✅ nextjs-turbopack-stable-lazy-discovery-enabled	129	0	0
✅ nextjs-webpack-canary	110	0	19
✅ nextjs-webpack-stable-lazy-discovery-disabled	129	0	0
✅ nextjs-webpack-stable-lazy-discovery-enabled	129	0	0
✅ nitro-stable	104	0	25
✅ nuxt-stable	104	0	25
✅ sveltekit-stable	123	0	6
✅ vite-stable	104	0	25

✅ 📦 Local Production

App	Passed	Failed	Skipped
✅ astro-stable	104	0	25
✅ express-stable	104	0	25
✅ fastify-stable	104	0	25
✅ hono-stable	104	0	25
✅ nextjs-turbopack-canary	110	0	19
✅ nextjs-turbopack-stable-lazy-discovery-disabled	129	0	0
✅ nextjs-webpack-canary	110	0	19
✅ nextjs-webpack-stable-lazy-discovery-disabled	129	0	0
✅ nextjs-webpack-stable-lazy-discovery-enabled	129	0	0
✅ nitro-stable	104	0	25
✅ nuxt-stable	104	0	25
✅ sveltekit-stable	123	0	6
✅ vite-stable	104	0	25

❌ 🐘 Local Postgres

App	Passed	Failed	Skipped
✅ astro-stable	104	0	25
✅ express-stable	104	0	25
✅ fastify-stable	104	0	25
✅ hono-stable	104	0	25
✅ nextjs-turbopack-canary	110	0	19
✅ nextjs-turbopack-stable-lazy-discovery-disabled	129	0	0
✅ nextjs-turbopack-stable-lazy-discovery-enabled	129	0	0
✅ nextjs-webpack-canary	110	0	19
❌ nextjs-webpack-stable-lazy-discovery-disabled	128	1	0
❌ nextjs-webpack-stable-lazy-discovery-enabled	128	1	0
✅ nitro-stable	104	0	25
✅ nuxt-stable	104	0	25
✅ sveltekit-stable	123	0	6
✅ vite-stable	104	0	25

✅ 🪟 Windows

App	Passed	Failed	Skipped
✅ nextjs-turbopack	129	0	0

✅ 📋 Other

App	Passed	Failed	Skipped
✅ e2e-local-dev-nest-stable	104	0	25
✅ e2e-local-dev-tanstack-start-	104	0	25
✅ e2e-local-postgres-nest-stable	104	0	25
✅ e2e-local-postgres-tanstack-start-	104	0	25
✅ e2e-local-prod-nest-stable	104	0	25
✅ e2e-local-prod-tanstack-start-	104	0	25
✅ e2e-vercel-prod-tanstack-start	103	0	26

---

📋 View full workflow run

---

❌ Some E2E test jobs failed:

• Vercel Prod: failure
• Local Dev: success
• Local Prod: failure
• Local Postgres: failure
• Windows: success

Check the workflow run for details.

[Copilot]

Pull request overview

Adds a new "Versioning" Foundations page (v4 and v5) explaining default run pinning to deployments, recovery via cancel + rerun, opting into newer code with deploymentId: "latest", and carrying serializable state/streams across child-run continuations. Existing references to deploymentId: "latest" and versioning behavior are linked back to the new page. Cookbook pages also pick up callouts about long-lived / multi-turn patterns and, on v5 only, are updated to reflect that start() can now be called directly inside "use workflow" functions (via #1491), while v4 retains the step-wrapped pattern.

Changes:

• New foundations/versioning.mdx for both v4 and v5, plus meta.json + index.mdx + linkbacks from vercel-world, start.mdx, and code-transform.mdx.
• v5 cookbook + migration guides updated to drop "use step" wrappers around start() in favor of calling start() directly from the workflow function; v4 keeps the step-wrapped pattern.
• Added "Versioning" callouts to scheduling, durable-agent, ai-sdk, chat-sdk, and sandbox cookbooks; fixed a few small wording/grammar issues in code-transform.mdx.

Reviewed changes

Copilot reviewed 32 out of 32 changed files in this pull request and generated no comments.

Show a summary per file

File	Description
docs/content/docs/v5/foundations/versioning.mdx	New Foundations page explaining run pinning, recovery, and explicit upgrade boundaries.
docs/content/docs/v4/foundations/versioning.mdx	v4 counterpart; uses step-wrapped continueDigest/continueExportOnLatest helpers.
docs/content/docs/v{4,5}/foundations/{meta.json,index.mdx}	Surface the new Versioning page in nav and the foundations index.
docs/content/docs/v{4,5}/deploying/world/vercel-world.mdx	Link the deployment-pinning section to the new Versioning page.
docs/content/docs/v{4,5}/how-it-works/code-transform.mdx	Link "atomic versioning" reference; fix "function assume" / "from being upgraded" wording.
docs/content/docs/v{4,5}/api-reference/workflow-api/start.mdx	Soften "Vercel-specific" to "Vercelism", mention possible version rename, link Versioning. v5 also notes native workflow-context start().
docs/content/docs/v5/foundations/starting-workflows.mdx	Note v5 can call start() from workflows; link Composition + Versioning.
docs/content/docs/v5/cookbook/common-patterns/workflow-composition.mdx	Replace step-wrapped child spawn with direct start() call.
docs/content/docs/v5/cookbook/advanced/child-workflows.mdx	Inline batched start() calls; rename spawnReportChunk → startReportChunk; update tips.
docs/content/docs/v{4,5}/cookbook/common-patterns/scheduling.mdx	Callout linking to Versioning for long-lived schedules.
docs/content/docs/v{4,5}/cookbook/agent-patterns/durable-agent.mdx	Callout linking to Versioning for multi-turn agents.
docs/content/docs/v{4,5}/cookbook/integrations/{ai-sdk,chat-sdk,sandbox}.mdx	Callouts linking to Versioning for long-lived sessions.
docs/content/docs/v4/cookbook/common-patterns/workflow-composition.mdx	Update deploymentId: "latest" callout wording + link to Versioning.
docs/content/docs/v4/cookbook/advanced/child-workflows.mdx	Link Versioning from the deploymentId: "latest" tip.
docs/content/docs/v4/changelog/eager-processing.mdx	Reword to "step-wrapped start() helper" to match v4 phrasing.
docs/content/docs/v5/migration-guides/{migrating-from-trigger-dev,migrating-from-temporal,migrating-from-inngest,migrating-from-aws-step-functions}.mdx	Update to v5 native workflow-context start(); only collect step retains "use step".

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[TooTallNate]

Comment-only review

The new versioning foundations page is a useful addition — the four sections (default pinning → recovery → self-upgrade via start({ deploymentId: "latest" }) → carrying serializable context forward) cover the model clearly, and the v4-vs-v5 differentiation (step-wrapped start() vs direct start() from workflow context, gated on whether PR #1491 has landed on the target branch) is correctly handled. The cross-links from cookbook patterns to the new foundations page tie everything together nicely.

Three observations — see inline.

Inline comments

1. Wording: this PR introduces "Vercelism" as a new term and uses it 6 times across the diff, including replacing existing "Vercel-specific feature" wording in two <Callout>s. "Vercelism" reads as informal slang and isn't part of the established docs vocabulary. The existing tone uses "Vercel-specific", "Vercel deployments", "Vercel World", "non-Vercel" — all precise. Suggested restoring the prior phrasing. This is the only thing I'd consider blocking.
2. Broken link in v4: versioning.mdx references /docs/foundations/cancellation in its related frontmatter, but that page is v5-only. Easy fix.
3. Missing changeset: convention for docs-only PRs is an empty ---\n--- file.

A few small things I noticed

• The pre-emptively check for a conflict with getHookByToken() example (line 96 of versioning.mdx) — nice that you noted the race window in a later doc (#2011 callout), but this page doesn't have that warning. Worth a cross-link or <Callout type="warn"> here too, since the recovery pattern uses both start() and deploymentId: "latest" together.
• The _idempotency / using syntax is great. Worth confirming the docs site's syntax highlighter renders it — I noticed #1988 recently fixed a homepage link lint regression, and the using syntax is TypeScript 5.2+ specific.
• dailyDigest example (line 137-152): the state parameter is DigestState but the workflow doesn't actually use state.userId after the first sendDigest call — pattern is fine, just noting if you wanted to make the example illustrate state evolution more vividly, having sendDigest read state.lastSentAt (e.g. to avoid sending duplicates) would make the carry-forward intent clearer.

Overall the page reads well — once the Vercelism wording is restored and the v4 link is fixed, this should be good to go.

[TooTallNate]

All three review comments addressed in 7eff96c — Vercelism wording reverted, v4 broken cancellation link fixed, empty changeset added. LGTM, approving.

[github-actions]

Backport PR opened against stable: #2014. Merge conflicts were resolved by AI — please review carefully.

[pranaygp]

Summary

Adds a new Foundations page explaining Workflow SDK versioning semantics for both v5 and v4 docs.

The page covers:

• default run pinning to the deployment that started the run
• canceling and rerunning affected in-flight runs on the latest deployment
• explicit self-upgrade patterns using deploymentId: "latest"
• carrying serializable context such as streams and abort signals across child runs

This also links existing versioning and deploymentId: "latest" mentions back to the new Foundations page, and calls out that deploymentId is currently a Vercelism that may become a more generic version concept in the World spec later.

Follow-up cookbook pass added targeted callouts for long-lived or multi-turn patterns:

• scheduled workflows
• DurableAgent
• AI SDK multi-turn sessions
• Chat SDK sessions
• Sandbox sessions

The docs now also distinguish v5 from v4 for child-run starts. Native start() from workflow context was added by e295bae41 / #1491, which is on main but not stable, so v5 examples encourage direct workflow-context start() and v4 keeps the step-wrapped pattern.

Validation

• git diff --check
• Parsed the updated v4/v5 Foundations meta.json files as JSON
• Verified v4/v5 copies of the touched cookbook pages match before the v5-only native-start pass
• Searched docs/content for stale v5 step-wrapped start() guidance and v4 direct-start guidance
• Confirmed origin/stable lacks the 'use step' directive in packages/core/src/runtime/start.ts, while main has it via e295bae41 / feat: allow start() to be called directly inside workflow functions #1491

Full docs lint/typecheck were not run because this checkout does not have node_modules installed, so biome, vitest, and changeset are unavailable locally.

[vercel]

The latest updates on your projects. Learn more about Vercel for GitHub.

Project	Deployment	Actions	Updated (UTC)
example-nextjs-workflow-turbopack	Ready	Preview, Comment	May 18, 2026 10:49pm
example-nextjs-workflow-webpack	Ready	Preview, Comment	May 18, 2026 10:49pm
example-workflow	Ready	Preview, Comment	May 18, 2026 10:49pm
workbench-astro-workflow	Ready	Preview, Comment	May 18, 2026 10:49pm
workbench-express-workflow	Ready	Preview, Comment	May 18, 2026 10:49pm
workbench-fastify-workflow	Ready	Preview, Comment	May 18, 2026 10:49pm
workbench-hono-workflow	Ready	Preview, Comment	May 18, 2026 10:49pm
workbench-nitro-workflow	Ready	Preview, Comment	May 18, 2026 10:49pm
workbench-nuxt-workflow	Ready	Preview, Comment	May 18, 2026 10:49pm
workbench-sveltekit-workflow	Ready	Preview, Comment	May 18, 2026 10:49pm
workbench-tanstack-start-workflow	Ready	Preview, Comment	May 18, 2026 10:49pm
workbench-vite-workflow	Ready	Preview, Comment	May 18, 2026 10:49pm
workflow-docs	Ready	Preview, ✅ 11 resolved, Open in v0	May 18, 2026 10:49pm
workflow-swc-playground	Ready	Preview, Comment	May 18, 2026 10:49pm
workflow-tarballs	Ready	Preview, Comment	May 18, 2026 10:49pm
workflow-web	Ready	Preview, Comment	May 18, 2026 10:49pm

[changeset-bot]

🦋 Changeset detected

Latest commit: 7eff96c

The changes in this PR will be included in the next version bump.

This PR includes changesets to release 0 packages

When changesets are added to this PR, you'll see the packages that this PR includes changesets for and the associated semver types

Not sure what this means? Click here to learn what changesets are.

Click here if you're a maintainer who wants to add another changeset to this PR

[github-actions]

📊 Benchmark Results

📈 Comparing against baseline from main branch. Green 🟢 = faster, Red 🔺 = slower.

workflow with no steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
💻 Local	🥇 Nitro	0.031s (-28.3% 🟢)	1.005s (~)	0.974s	10	1.00x
💻 Local	Express	0.036s (-19.6% 🟢)	1.005s (~)	0.970s	10	1.15x
🐘 Postgres	Express	0.045s (-21.9% 🟢)	1.011s (~)	0.966s	10	1.47x
🐘 Postgres	Nitro	0.046s (-51.3% 🟢)	1.013s (-2.9%)	0.967s	10	1.50x

workflow with 1 step

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
💻 Local	🥇 Express	1.070s (-4.9%)	2.006s (~)	0.936s	10	1.00x
💻 Local	Nitro	1.071s (-5.3% 🟢)	2.006s (~)	0.935s	10	1.00x
🐘 Postgres	Express	1.079s (-5.9% 🟢)	2.009s (~)	0.930s	10	1.01x
🐘 Postgres	Nitro	1.081s (-5.2% 🟢)	2.010s (~)	0.929s	10	1.01x

workflow with 10 sequential steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Nitro	10.396s (-4.4%)	11.018s (~)	0.622s	3	1.00x
🐘 Postgres	Express	10.402s (-5.1% 🟢)	11.016s (~)	0.614s	3	1.00x
💻 Local	Express	10.411s (-4.7%)	11.023s (~)	0.612s	3	1.00x
💻 Local	Nitro	10.412s (-4.9%)	11.022s (~)	0.610s	3	1.00x

workflow with 25 sequential steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
💻 Local	🥇 Express	13.432s (-10.3% 🟢)	14.027s (-6.7% 🟢)	0.595s	5	1.00x
🐘 Postgres	Express	13.435s (-7.9% 🟢)	14.017s (-6.7% 🟢)	0.581s	5	1.00x
🐘 Postgres	Nitro	13.484s (-7.6% 🟢)	14.017s (-6.7% 🟢)	0.533s	5	1.00x
💻 Local	Nitro	13.489s (-10.4% 🟢)	14.027s (-12.5% 🟢)	0.538s	5	1.00x

workflow with 50 sequential steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
💻 Local	🥇 Nitro	11.875s (-29.2% 🟢)	12.022s (-29.4% 🟢)	0.147s	8	1.00x
💻 Local	Express	11.890s (-28.4% 🟢)	12.022s (-29.4% 🟢)	0.132s	8	1.00x
🐘 Postgres	Express	11.941s (-14.7% 🟢)	12.391s (-15.1% 🟢)	0.450s	8	1.01x
🐘 Postgres	Nitro	11.944s (-14.5% 🟢)	12.266s (-14.3% 🟢)	0.321s	8	1.01x

Promise.all with 10 concurrent steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Express	1.136s (-9.8% 🟢)	2.007s (~)	0.870s	15	1.00x
🐘 Postgres	Nitro	1.137s (-10.8% 🟢)	2.007s (~)	0.870s	15	1.00x
💻 Local	Nitro	1.172s (-28.2% 🟢)	2.006s (-3.3%)	0.834s	15	1.03x
💻 Local	Express	1.183s (-20.6% 🟢)	2.006s (~)	0.823s	15	1.04x

Promise.all with 25 concurrent steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Nitro	1.192s (-49.3% 🟢)	2.007s (-33.3% 🟢)	0.815s	15	1.00x
🐘 Postgres	Express	1.224s (-48.2% 🟢)	2.007s (-33.3% 🟢)	0.783s	15	1.03x
💻 Local	Express	1.703s (-42.3% 🟢)	2.005s (-41.9% 🟢)	0.303s	15	1.43x
💻 Local	Nitro	1.704s (-45.8% 🟢)	2.005s (-48.4% 🟢)	0.302s	15	1.43x

Promise.all with 50 concurrent steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Nitro	1.305s (-62.5% 🟢)	2.008s (-49.9% 🟢)	0.702s	15	1.00x
🐘 Postgres	Express	1.324s (-62.0% 🟢)	2.008s (-49.9% 🟢)	0.684s	15	1.01x
💻 Local	Express	4.638s (-44.4% 🟢)	5.179s (-42.6% 🟢)	0.541s	6	3.55x
💻 Local	Nitro	4.910s (-41.2% 🟢)	5.511s (-38.9% 🟢)	0.601s	6	3.76x

Promise.race with 10 concurrent steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Express	1.155s (-8.1% 🟢)	2.009s (~)	0.854s	15	1.00x
🐘 Postgres	Nitro	1.163s (-7.5% 🟢)	2.010s (~)	0.847s	15	1.01x
💻 Local	Express	1.386s (-26.8% 🟢)	2.006s (-15.2% 🟢)	0.620s	15	1.20x
💻 Local	Nitro	1.388s (-25.6% 🟢)	2.006s (-14.3% 🟢)	0.618s	15	1.20x

Promise.race with 25 concurrent steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Express	1.199s (-48.8% 🟢)	2.008s (-33.3% 🟢)	0.810s	15	1.00x
🐘 Postgres	Nitro	1.199s (-48.7% 🟢)	2.007s (-33.3% 🟢)	0.808s	15	1.00x
💻 Local	Express	1.828s (-41.6% 🟢)	2.222s (-40.9% 🟢)	0.394s	14	1.53x
💻 Local	Nitro	1.935s (-36.9% 🟢)	2.316s (-40.4% 🟢)	0.381s	13	1.61x

Promise.race with 50 concurrent steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Express	1.305s (-62.7% 🟢)	2.008s (-49.9% 🟢)	0.702s	15	1.00x
🐘 Postgres	Nitro	1.309s (-62.4% 🟢)	2.009s (-49.9% 🟢)	0.700s	15	1.00x
💻 Local	Express	5.446s (-38.1% 🟢)	5.849s (-36.9% 🟢)	0.403s	6	4.17x
💻 Local	Nitro	5.557s (-39.2% 🟢)	6.179s (-38.4% 🟢)	0.622s	6	4.26x

workflow with 10 sequential data payload steps (10KB)

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Express	0.418s (-50.1% 🟢)	1.006s (-1.6%)	0.588s	60	1.00x
🐘 Postgres	Nitro	0.459s (-44.0% 🟢)	1.006s (~)	0.547s	60	1.10x
💻 Local	Nitro	0.468s (-52.2% 🟢)	1.004s (-8.2% 🟢)	0.535s	60	1.12x
💻 Local	Express	0.469s (-52.4% 🟢)	1.004s (-6.7% 🟢)	0.535s	60	1.12x

workflow with 25 sequential data payload steps (10KB)

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Express	1.047s (-47.0% 🟢)	1.705s (-24.5% 🟢)	0.659s	53	1.00x
🐘 Postgres	Nitro	1.066s (-44.7% 🟢)	1.576s (-25.0% 🟢)	0.510s	58	1.02x
💻 Local	Nitro	1.185s (-61.0% 🟢)	2.006s (-46.6% 🟢)	0.821s	45	1.13x
💻 Local	Express	1.191s (-60.5% 🟢)	2.005s (-44.1% 🟢)	0.815s	45	1.14x

workflow with 50 sequential data payload steps (10KB)

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Express	2.030s (-49.1% 🟢)	2.427s (-44.4% 🟢)	0.397s	50	1.00x
🐘 Postgres	Nitro	2.128s (-48.2% 🟢)	2.660s (-42.2% 🟢)	0.532s	46	1.05x
💻 Local	Nitro	2.704s (-70.9% 🟢)	3.032s (-69.7% 🟢)	0.328s	40	1.33x
💻 Local	Express	2.719s (-70.5% 🟢)	3.057s (-69.5% 🟢)	0.338s	40	1.34x

workflow with 10 concurrent data payload steps (10KB)

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Express	0.163s (-42.4% 🟢)	1.005s (~)	0.843s	60	1.00x
🐘 Postgres	Nitro	0.178s (-37.0% 🟢)	1.006s (~)	0.827s	60	1.10x
💻 Local	Express	0.455s (-18.8% 🟢)	1.004s (~)	0.550s	60	2.80x
💻 Local	Nitro	0.469s (-22.4% 🟢)	1.004s (-1.7%)	0.535s	60	2.89x

workflow with 25 concurrent data payload steps (10KB)

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Express	0.286s (-43.9% 🟢)	1.006s (~)	0.720s	90	1.00x
🐘 Postgres	Nitro	0.295s (-40.5% 🟢)	1.006s (~)	0.711s	90	1.03x
💻 Local	Nitro	2.205s (-13.1% 🟢)	2.944s (-2.2%)	0.738s	31	7.71x
💻 Local	Express	2.226s (-11.4% 🟢)	2.943s (-2.2%)	0.717s	31	7.78x

workflow with 50 concurrent data payload steps (10KB)

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Express	0.582s (-28.9% 🟢)	1.007s (-1.0%)	0.425s	120	1.00x
🐘 Postgres	Nitro	0.634s (-19.7% 🟢)	1.006s (~)	0.372s	120	1.09x
💻 Local	Nitro	10.021s (-10.5% 🟢)	10.611s (-9.0% 🟢)	0.590s	12	17.21x
💻 Local	Express	10.042s (-10.3% 🟢)	10.694s (-10.4% 🟢)	0.652s	12	17.25x

Stream Benchmarks(includes TTFB metrics)

workflow with stream

💻 Local Development

World	Framework	Workflow Time	TTFB	Slurp	Wall Time	Overhead	Samples	vs Fastest
💻 Local	🥇 Express	1.132s (+468.5% 🔺)	2.005s (+99.6% 🔺)	0.011s (-13.2% 🟢)	2.018s (+98.2% 🔺)	0.886s	10	1.00x
💻 Local	Nitro	1.133s (+430.4% 🔺)	2.005s (+99.6% 🔺)	0.011s (-15.2% 🟢)	2.018s (+98.1% 🔺)	0.885s	10	1.00x
🐘 Postgres	Nitro	1.135s (+453.6% 🔺)	2.001s (+100.2% 🔺)	0.001s (-20.0% 🟢)	2.010s (+98.7% 🔺)	0.875s	10	1.00x
🐘 Postgres	Express	1.135s (+453.5% 🔺)	1.999s (+100.2% 🔺)	0.001s (-25.0% 🟢)	2.010s (+98.7% 🔺)	0.875s	10	1.00x

stream pipeline with 5 transform steps (1MB)

💻 Local Development

World	Framework	Workflow Time	TTFB	Slurp	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Express	1.492s (+136.8% 🔺)	2.002s (+98.9% 🔺)	0.003s (-30.4% 🟢)	2.023s (+97.8% 🔺)	0.531s	30	1.00x
🐘 Postgres	Nitro	1.519s (+143.4% 🔺)	2.004s (+99.0% 🔺)	0.004s (-2.5%)	2.027s (+98.2% 🔺)	0.507s	30	1.02x
💻 Local	Nitro	1.707s (+103.5% 🔺)	2.013s (+98.9% 🔺)	0.009s (-4.3%)	2.203s (+97.4% 🔺)	0.496s	28	1.14x
💻 Local	Express	1.923s (+154.0% 🔺)	2.011s (+95.4% 🔺)	0.009s (-8.2% 🟢)	2.422s (+132.9% 🔺)	0.499s	25	1.29x

10 parallel streams (1MB each)

💻 Local Development

World	Framework	Workflow Time	TTFB	Slurp	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Express	0.627s (-34.7% 🟢)	1.033s (-19.2% 🟢)	0.000s (-20.7% 🟢)	1.045s (-20.0% 🟢)	0.418s	58	1.00x
🐘 Postgres	Nitro	0.662s (-31.7% 🟢)	1.029s (-17.5% 🟢)	0.000s (-17.2% 🟢)	1.049s (-16.6% 🟢)	0.387s	58	1.05x
💻 Local	Express	1.365s (+11.5% 🔺)	2.015s (~)	0.000s (+40.0% 🔺)	2.018s (~)	0.652s	30	2.18x
💻 Local	Nitro	1.561s (+27.6% 🔺)	2.016s (~)	0.000s (+364.3% 🔺)	2.197s (+8.7% 🔺)	0.636s	28	2.49x

fan-out fan-in 10 streams (1MB each)

💻 Local Development

World	Framework	Workflow Time	TTFB	Slurp	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Express	1.228s (-30.7% 🟢)	1.869s (-14.2% 🟢)	0.000s (NaN%)	1.900s (-13.6% 🟢)	0.673s	32	1.00x
🐘 Postgres	Nitro	1.325s (-26.1% 🟢)	1.997s (-6.7% 🟢)	0.000s (+86.7% 🔺)	2.019s (-7.2% 🟢)	0.694s	30	1.08x
💻 Local	Express	3.158s (-8.9% 🟢)	4.028s (~)	0.001s (-33.3% 🟢)	4.030s (~)	0.873s	15	2.57x
💻 Local	Nitro	3.187s (-5.9% 🟢)	4.029s (~)	0.000s (-75.0% 🟢)	4.031s (~)	0.844s	15	2.60x

Summary

Fastest Framework by World

Winner determined by most benchmark wins

World	🥇 Fastest Framework	Wins
💻 Local	Express	12/21
🐘 Postgres	Express	17/21

Fastest World by Framework

Winner determined by most benchmark wins

Framework	🥇 Fastest World	Wins
Express	🐘 Postgres	16/21
Nitro	🐘 Postgres	17/21

Column Definitions

• Workflow Time: Runtime reported by workflow (completedAt - createdAt) - primary metric
• TTFB: Time to First Byte - time from workflow start until first stream byte received (stream benchmarks only)
• Slurp: Time from first byte to complete stream consumption (stream benchmarks only)
• Wall Time: Total testbench time (trigger workflow + poll for result)
• Overhead: Testbench overhead (Wall Time - Workflow Time)
• Samples: Number of benchmark iterations run
• vs Fastest: How much slower compared to the fastest configuration for this benchmark

Worlds:

• 💻 Local: In-memory filesystem world (local development)
• 🐘 Postgres: PostgreSQL database world (local development)
• ▲ Vercel: Vercel production/preview deployment
• 🌐 Turso: Community world (local development)
• 🌐 MongoDB: Community world (local development)
• 🌐 Redis: Community world (local development)
• 🌐 Jazz: Community world (local development)
• 🌐 Redis: Community world (local development)
• 🌐 Redis + BullMQ: Community world (local development)
• 🌐 Cloudflare: Community world (local development)
• 🌐 MySQL: Community world (local development)
• 🌐 Azure: Community world (local development)
• 🌐 NATS JetStream: Community world (local development)
• 🌐 Upstash: Community world (local development)

---

📋 View full workflow run

---

❌ Some benchmark jobs failed:

• Local: success
• Postgres: success
• Vercel: failure

Check the workflow run for details.

[github-actions]

🧪 E2E Test Results

❌ Some tests failed

Summary

	Passed	Failed	Skipped	Total
❌ ▲ Vercel Production	1196	4	219	1419
✅ 💻 Local Development	1587	0	219	1806
✅ 📦 Local Production	1458	0	219	1677
❌ 🐘 Local Postgres	1585	2	219	1806
✅ 🪟 Windows	129	0	0	129
✅ 📋 Other	727	0	176	903
Total	6682	6	1052	7740

❌ Failed Tests

▲ Vercel Production (4 failed)

astro (1 failed):

• AbortController abortFromStepWorkflow: step abort cancels an in-flight sibling step

hono (2 failed):

• outputStreamWorkflow no startIndex (reads all chunks)
• stepFunctionPassingWorkflow - step function references can be passed as arguments (without closure vars) | wrun_01KRYMPWTNKF0CE4BP27X7R946 | 🔍 observability

nextjs-webpack (1 failed):

• parallelSleepWorkflow | wrun_01KRYME5H0F7D8EH2AAF56T0Z2 | 🔍 observability

🐘 Local Postgres (2 failed)

nextjs-webpack-stable-lazy-discovery-disabled (1 failed):

• DurableAgent e2e core basic text response

nextjs-webpack-stable-lazy-discovery-enabled (1 failed):

• addTenWorkflow | wrun_01KRYMBT8QHFK9EZ9FDY03D9G0

Details by Category

❌ ▲ Vercel Production

App	Passed	Failed	Skipped
❌ astro	102	1	26
✅ example	103	0	26
✅ express	103	0	26
✅ fastify	103	0	26
❌ hono	101	2	26
✅ nextjs-turbopack	127	0	2
❌ nextjs-webpack	126	1	2
✅ nitro	103	0	26
✅ nuxt	103	0	26
✅ sveltekit	122	0	7
✅ vite	103	0	26

✅ 💻 Local Development

App	Passed	Failed	Skipped
✅ astro-stable	104	0	25
✅ express-stable	104	0	25
✅ fastify-stable	104	0	25
✅ hono-stable	104	0	25
✅ nextjs-turbopack-canary	110	0	19
✅ nextjs-turbopack-stable-lazy-discovery-disabled	129	0	0
✅ nextjs-turbopack-stable-lazy-discovery-enabled	129	0	0
✅ nextjs-webpack-canary	110	0	19
✅ nextjs-webpack-stable-lazy-discovery-disabled	129	0	0
✅ nextjs-webpack-stable-lazy-discovery-enabled	129	0	0
✅ nitro-stable	104	0	25
✅ nuxt-stable	104	0	25
✅ sveltekit-stable	123	0	6
✅ vite-stable	104	0	25

✅ 📦 Local Production

App	Passed	Failed	Skipped
✅ astro-stable	104	0	25
✅ express-stable	104	0	25
✅ fastify-stable	104	0	25
✅ hono-stable	104	0	25
✅ nextjs-turbopack-canary	110	0	19
✅ nextjs-turbopack-stable-lazy-discovery-disabled	129	0	0
✅ nextjs-webpack-canary	110	0	19
✅ nextjs-webpack-stable-lazy-discovery-disabled	129	0	0
✅ nextjs-webpack-stable-lazy-discovery-enabled	129	0	0
✅ nitro-stable	104	0	25
✅ nuxt-stable	104	0	25
✅ sveltekit-stable	123	0	6
✅ vite-stable	104	0	25

❌ 🐘 Local Postgres

App	Passed	Failed	Skipped
✅ astro-stable	104	0	25
✅ express-stable	104	0	25
✅ fastify-stable	104	0	25
✅ hono-stable	104	0	25
✅ nextjs-turbopack-canary	110	0	19
✅ nextjs-turbopack-stable-lazy-discovery-disabled	129	0	0
✅ nextjs-turbopack-stable-lazy-discovery-enabled	129	0	0
✅ nextjs-webpack-canary	110	0	19
❌ nextjs-webpack-stable-lazy-discovery-disabled	128	1	0
❌ nextjs-webpack-stable-lazy-discovery-enabled	128	1	0
✅ nitro-stable	104	0	25
✅ nuxt-stable	104	0	25
✅ sveltekit-stable	123	0	6
✅ vite-stable	104	0	25

✅ 🪟 Windows

App	Passed	Failed	Skipped
✅ nextjs-turbopack	129	0	0

✅ 📋 Other

App	Passed	Failed	Skipped
✅ e2e-local-dev-nest-stable	104	0	25
✅ e2e-local-dev-tanstack-start-	104	0	25
✅ e2e-local-postgres-nest-stable	104	0	25
✅ e2e-local-postgres-tanstack-start-	104	0	25
✅ e2e-local-prod-nest-stable	104	0	25
✅ e2e-local-prod-tanstack-start-	104	0	25
✅ e2e-vercel-prod-tanstack-start	103	0	26

---

📋 View full workflow run

---

❌ Some E2E test jobs failed:

• Vercel Prod: failure
• Local Dev: success
• Local Prod: failure
• Local Postgres: failure
• Windows: success

Check the workflow run for details.

[Copilot]

Pull request overview

Adds a new "Versioning" Foundations page (v4 and v5) explaining default run pinning to deployments, recovery via cancel + rerun, opting into newer code with deploymentId: "latest", and carrying serializable state/streams across child-run continuations. Existing references to deploymentId: "latest" and versioning behavior are linked back to the new page. Cookbook pages also pick up callouts about long-lived / multi-turn patterns and, on v5 only, are updated to reflect that start() can now be called directly inside "use workflow" functions (via #1491), while v4 retains the step-wrapped pattern.

Changes:

• New foundations/versioning.mdx for both v4 and v5, plus meta.json + index.mdx + linkbacks from vercel-world, start.mdx, and code-transform.mdx.
• v5 cookbook + migration guides updated to drop "use step" wrappers around start() in favor of calling start() directly from the workflow function; v4 keeps the step-wrapped pattern.
• Added "Versioning" callouts to scheduling, durable-agent, ai-sdk, chat-sdk, and sandbox cookbooks; fixed a few small wording/grammar issues in code-transform.mdx.

Reviewed changes

Copilot reviewed 32 out of 32 changed files in this pull request and generated no comments.

Show a summary per file

File	Description
docs/content/docs/v5/foundations/versioning.mdx	New Foundations page explaining run pinning, recovery, and explicit upgrade boundaries.
docs/content/docs/v4/foundations/versioning.mdx	v4 counterpart; uses step-wrapped continueDigest/continueExportOnLatest helpers.
docs/content/docs/v{4,5}/foundations/{meta.json,index.mdx}	Surface the new Versioning page in nav and the foundations index.
docs/content/docs/v{4,5}/deploying/world/vercel-world.mdx	Link the deployment-pinning section to the new Versioning page.
docs/content/docs/v{4,5}/how-it-works/code-transform.mdx	Link "atomic versioning" reference; fix "function assume" / "from being upgraded" wording.
docs/content/docs/v{4,5}/api-reference/workflow-api/start.mdx	Soften "Vercel-specific" to "Vercelism", mention possible version rename, link Versioning. v5 also notes native workflow-context start().
docs/content/docs/v5/foundations/starting-workflows.mdx	Note v5 can call start() from workflows; link Composition + Versioning.
docs/content/docs/v5/cookbook/common-patterns/workflow-composition.mdx	Replace step-wrapped child spawn with direct start() call.
docs/content/docs/v5/cookbook/advanced/child-workflows.mdx	Inline batched start() calls; rename spawnReportChunk → startReportChunk; update tips.
docs/content/docs/v{4,5}/cookbook/common-patterns/scheduling.mdx	Callout linking to Versioning for long-lived schedules.
docs/content/docs/v{4,5}/cookbook/agent-patterns/durable-agent.mdx	Callout linking to Versioning for multi-turn agents.
docs/content/docs/v{4,5}/cookbook/integrations/{ai-sdk,chat-sdk,sandbox}.mdx	Callouts linking to Versioning for long-lived sessions.
docs/content/docs/v4/cookbook/common-patterns/workflow-composition.mdx	Update deploymentId: "latest" callout wording + link to Versioning.
docs/content/docs/v4/cookbook/advanced/child-workflows.mdx	Link Versioning from the deploymentId: "latest" tip.
docs/content/docs/v4/changelog/eager-processing.mdx	Reword to "step-wrapped start() helper" to match v4 phrasing.
docs/content/docs/v5/migration-guides/{migrating-from-trigger-dev,migrating-from-temporal,migrating-from-inngest,migrating-from-aws-step-functions}.mdx	Update to v5 native workflow-context start(); only collect step retains "use step".

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[TooTallNate]

Comment-only review

The new versioning foundations page is a useful addition — the four sections (default pinning → recovery → self-upgrade via start({ deploymentId: "latest" }) → carrying serializable context forward) cover the model clearly, and the v4-vs-v5 differentiation (step-wrapped start() vs direct start() from workflow context, gated on whether PR #1491 has landed on the target branch) is correctly handled. The cross-links from cookbook patterns to the new foundations page tie everything together nicely.

Three observations — see inline.

Inline comments

1. Wording: this PR introduces "Vercelism" as a new term and uses it 6 times across the diff, including replacing existing "Vercel-specific feature" wording in two <Callout>s. "Vercelism" reads as informal slang and isn't part of the established docs vocabulary. The existing tone uses "Vercel-specific", "Vercel deployments", "Vercel World", "non-Vercel" — all precise. Suggested restoring the prior phrasing. This is the only thing I'd consider blocking.
2. Broken link in v4: versioning.mdx references /docs/foundations/cancellation in its related frontmatter, but that page is v5-only. Easy fix.
3. Missing changeset: convention for docs-only PRs is an empty ---\n--- file.

A few small things I noticed

• The pre-emptively check for a conflict with getHookByToken() example (line 96 of versioning.mdx) — nice that you noted the race window in a later doc (#2011 callout), but this page doesn't have that warning. Worth a cross-link or <Callout type="warn"> here too, since the recovery pattern uses both start() and deploymentId: "latest" together.
• The _idempotency / using syntax is great. Worth confirming the docs site's syntax highlighter renders it — I noticed #1988 recently fixed a homepage link lint regression, and the using syntax is TypeScript 5.2+ specific.
• dailyDigest example (line 137-152): the state parameter is DigestState but the workflow doesn't actually use state.userId after the first sendDigest call — pattern is fine, just noting if you wanted to make the example illustrate state evolution more vividly, having sendDigest read state.lastSentAt (e.g. to avoid sending duplicates) would make the carry-forward intent clearer.

Overall the page reads well — once the Vercelism wording is restored and the v4 link is fixed, this should be good to go.

[TooTallNate]

All three review comments addressed in 7eff96c — Vercelism wording reverted, v4 broken cancellation link fixed, empty changeset added. LGTM, approving.

[github-actions]

Backport PR opened against stable: #2014. Merge conflicts were resolved by AI — please review carefully.

[pranaygp]

Summary

Adds a new Foundations page explaining Workflow SDK versioning semantics for both v5 and v4 docs.

The page covers:

• default run pinning to the deployment that started the run
• canceling and rerunning affected in-flight runs on the latest deployment
• explicit self-upgrade patterns using deploymentId: "latest"
• carrying serializable context such as streams and abort signals across child runs

This also links existing versioning and deploymentId: "latest" mentions back to the new Foundations page, and calls out that deploymentId is currently a Vercelism that may become a more generic version concept in the World spec later.

Follow-up cookbook pass added targeted callouts for long-lived or multi-turn patterns:

• scheduled workflows
• DurableAgent
• AI SDK multi-turn sessions
• Chat SDK sessions
• Sandbox sessions

The docs now also distinguish v5 from v4 for child-run starts. Native start() from workflow context was added by e295bae41 / #1491, which is on main but not stable, so v5 examples encourage direct workflow-context start() and v4 keeps the step-wrapped pattern.

Validation

• git diff --check
• Parsed the updated v4/v5 Foundations meta.json files as JSON
• Verified v4/v5 copies of the touched cookbook pages match before the v5-only native-start pass
• Searched docs/content for stale v5 step-wrapped start() guidance and v4 direct-start guidance
• Confirmed origin/stable lacks the 'use step' directive in packages/core/src/runtime/start.ts, while main has it via e295bae41 / feat: allow start() to be called directly inside workflow functions #1491

Full docs lint/typecheck were not run because this checkout does not have node_modules installed, so biome, vitest, and changeset are unavailable locally.

[vercel]

The latest updates on your projects. Learn more about Vercel for GitHub.

Project	Deployment	Actions	Updated (UTC)
example-nextjs-workflow-turbopack	Ready	Preview, Comment	May 18, 2026 10:49pm
example-nextjs-workflow-webpack	Ready	Preview, Comment	May 18, 2026 10:49pm
example-workflow	Ready	Preview, Comment	May 18, 2026 10:49pm
workbench-astro-workflow	Ready	Preview, Comment	May 18, 2026 10:49pm
workbench-express-workflow	Ready	Preview, Comment	May 18, 2026 10:49pm
workbench-fastify-workflow	Ready	Preview, Comment	May 18, 2026 10:49pm
workbench-hono-workflow	Ready	Preview, Comment	May 18, 2026 10:49pm
workbench-nitro-workflow	Ready	Preview, Comment	May 18, 2026 10:49pm
workbench-nuxt-workflow	Ready	Preview, Comment	May 18, 2026 10:49pm
workbench-sveltekit-workflow	Ready	Preview, Comment	May 18, 2026 10:49pm
workbench-tanstack-start-workflow	Ready	Preview, Comment	May 18, 2026 10:49pm
workbench-vite-workflow	Ready	Preview, Comment	May 18, 2026 10:49pm
workflow-docs	Ready	Preview, ✅ 11 resolved, Open in v0	May 18, 2026 10:49pm
workflow-swc-playground	Ready	Preview, Comment	May 18, 2026 10:49pm
workflow-tarballs	Ready	Preview, Comment	May 18, 2026 10:49pm
workflow-web	Ready	Preview, Comment	May 18, 2026 10:49pm

[changeset-bot]

🦋 Changeset detected

Latest commit: 7eff96c

The changes in this PR will be included in the next version bump.

This PR includes changesets to release 0 packages

When changesets are added to this PR, you'll see the packages that this PR includes changesets for and the associated semver types

Not sure what this means? Click here to learn what changesets are.

Click here if you're a maintainer who wants to add another changeset to this PR

[github-actions]

📊 Benchmark Results

📈 Comparing against baseline from main branch. Green 🟢 = faster, Red 🔺 = slower.

workflow with no steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
💻 Local	🥇 Nitro	0.031s (-28.3% 🟢)	1.005s (~)	0.974s	10	1.00x
💻 Local	Express	0.036s (-19.6% 🟢)	1.005s (~)	0.970s	10	1.15x
🐘 Postgres	Express	0.045s (-21.9% 🟢)	1.011s (~)	0.966s	10	1.47x
🐘 Postgres	Nitro	0.046s (-51.3% 🟢)	1.013s (-2.9%)	0.967s	10	1.50x

workflow with 1 step

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
💻 Local	🥇 Express	1.070s (-4.9%)	2.006s (~)	0.936s	10	1.00x
💻 Local	Nitro	1.071s (-5.3% 🟢)	2.006s (~)	0.935s	10	1.00x
🐘 Postgres	Express	1.079s (-5.9% 🟢)	2.009s (~)	0.930s	10	1.01x
🐘 Postgres	Nitro	1.081s (-5.2% 🟢)	2.010s (~)	0.929s	10	1.01x

workflow with 10 sequential steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Nitro	10.396s (-4.4%)	11.018s (~)	0.622s	3	1.00x
🐘 Postgres	Express	10.402s (-5.1% 🟢)	11.016s (~)	0.614s	3	1.00x
💻 Local	Express	10.411s (-4.7%)	11.023s (~)	0.612s	3	1.00x
💻 Local	Nitro	10.412s (-4.9%)	11.022s (~)	0.610s	3	1.00x

workflow with 25 sequential steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
💻 Local	🥇 Express	13.432s (-10.3% 🟢)	14.027s (-6.7% 🟢)	0.595s	5	1.00x
🐘 Postgres	Express	13.435s (-7.9% 🟢)	14.017s (-6.7% 🟢)	0.581s	5	1.00x
🐘 Postgres	Nitro	13.484s (-7.6% 🟢)	14.017s (-6.7% 🟢)	0.533s	5	1.00x
💻 Local	Nitro	13.489s (-10.4% 🟢)	14.027s (-12.5% 🟢)	0.538s	5	1.00x

workflow with 50 sequential steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
💻 Local	🥇 Nitro	11.875s (-29.2% 🟢)	12.022s (-29.4% 🟢)	0.147s	8	1.00x
💻 Local	Express	11.890s (-28.4% 🟢)	12.022s (-29.4% 🟢)	0.132s	8	1.00x
🐘 Postgres	Express	11.941s (-14.7% 🟢)	12.391s (-15.1% 🟢)	0.450s	8	1.01x
🐘 Postgres	Nitro	11.944s (-14.5% 🟢)	12.266s (-14.3% 🟢)	0.321s	8	1.01x

Promise.all with 10 concurrent steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Express	1.136s (-9.8% 🟢)	2.007s (~)	0.870s	15	1.00x
🐘 Postgres	Nitro	1.137s (-10.8% 🟢)	2.007s (~)	0.870s	15	1.00x
💻 Local	Nitro	1.172s (-28.2% 🟢)	2.006s (-3.3%)	0.834s	15	1.03x
💻 Local	Express	1.183s (-20.6% 🟢)	2.006s (~)	0.823s	15	1.04x

Promise.all with 25 concurrent steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Nitro	1.192s (-49.3% 🟢)	2.007s (-33.3% 🟢)	0.815s	15	1.00x
🐘 Postgres	Express	1.224s (-48.2% 🟢)	2.007s (-33.3% 🟢)	0.783s	15	1.03x
💻 Local	Express	1.703s (-42.3% 🟢)	2.005s (-41.9% 🟢)	0.303s	15	1.43x
💻 Local	Nitro	1.704s (-45.8% 🟢)	2.005s (-48.4% 🟢)	0.302s	15	1.43x

Promise.all with 50 concurrent steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Nitro	1.305s (-62.5% 🟢)	2.008s (-49.9% 🟢)	0.702s	15	1.00x
🐘 Postgres	Express	1.324s (-62.0% 🟢)	2.008s (-49.9% 🟢)	0.684s	15	1.01x
💻 Local	Express	4.638s (-44.4% 🟢)	5.179s (-42.6% 🟢)	0.541s	6	3.55x
💻 Local	Nitro	4.910s (-41.2% 🟢)	5.511s (-38.9% 🟢)	0.601s	6	3.76x

Promise.race with 10 concurrent steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Express	1.155s (-8.1% 🟢)	2.009s (~)	0.854s	15	1.00x
🐘 Postgres	Nitro	1.163s (-7.5% 🟢)	2.010s (~)	0.847s	15	1.01x
💻 Local	Express	1.386s (-26.8% 🟢)	2.006s (-15.2% 🟢)	0.620s	15	1.20x
💻 Local	Nitro	1.388s (-25.6% 🟢)	2.006s (-14.3% 🟢)	0.618s	15	1.20x

Promise.race with 25 concurrent steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Express	1.199s (-48.8% 🟢)	2.008s (-33.3% 🟢)	0.810s	15	1.00x
🐘 Postgres	Nitro	1.199s (-48.7% 🟢)	2.007s (-33.3% 🟢)	0.808s	15	1.00x
💻 Local	Express	1.828s (-41.6% 🟢)	2.222s (-40.9% 🟢)	0.394s	14	1.53x
💻 Local	Nitro	1.935s (-36.9% 🟢)	2.316s (-40.4% 🟢)	0.381s	13	1.61x

Promise.race with 50 concurrent steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Express	1.305s (-62.7% 🟢)	2.008s (-49.9% 🟢)	0.702s	15	1.00x
🐘 Postgres	Nitro	1.309s (-62.4% 🟢)	2.009s (-49.9% 🟢)	0.700s	15	1.00x
💻 Local	Express	5.446s (-38.1% 🟢)	5.849s (-36.9% 🟢)	0.403s	6	4.17x
💻 Local	Nitro	5.557s (-39.2% 🟢)	6.179s (-38.4% 🟢)	0.622s	6	4.26x

workflow with 10 sequential data payload steps (10KB)

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Express	0.418s (-50.1% 🟢)	1.006s (-1.6%)	0.588s	60	1.00x
🐘 Postgres	Nitro	0.459s (-44.0% 🟢)	1.006s (~)	0.547s	60	1.10x
💻 Local	Nitro	0.468s (-52.2% 🟢)	1.004s (-8.2% 🟢)	0.535s	60	1.12x
💻 Local	Express	0.469s (-52.4% 🟢)	1.004s (-6.7% 🟢)	0.535s	60	1.12x

workflow with 25 sequential data payload steps (10KB)

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Express	1.047s (-47.0% 🟢)	1.705s (-24.5% 🟢)	0.659s	53	1.00x
🐘 Postgres	Nitro	1.066s (-44.7% 🟢)	1.576s (-25.0% 🟢)	0.510s	58	1.02x
💻 Local	Nitro	1.185s (-61.0% 🟢)	2.006s (-46.6% 🟢)	0.821s	45	1.13x
💻 Local	Express	1.191s (-60.5% 🟢)	2.005s (-44.1% 🟢)	0.815s	45	1.14x

workflow with 50 sequential data payload steps (10KB)

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Express	2.030s (-49.1% 🟢)	2.427s (-44.4% 🟢)	0.397s	50	1.00x
🐘 Postgres	Nitro	2.128s (-48.2% 🟢)	2.660s (-42.2% 🟢)	0.532s	46	1.05x
💻 Local	Nitro	2.704s (-70.9% 🟢)	3.032s (-69.7% 🟢)	0.328s	40	1.33x
💻 Local	Express	2.719s (-70.5% 🟢)	3.057s (-69.5% 🟢)	0.338s	40	1.34x

workflow with 10 concurrent data payload steps (10KB)

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Express	0.163s (-42.4% 🟢)	1.005s (~)	0.843s	60	1.00x
🐘 Postgres	Nitro	0.178s (-37.0% 🟢)	1.006s (~)	0.827s	60	1.10x
💻 Local	Express	0.455s (-18.8% 🟢)	1.004s (~)	0.550s	60	2.80x
💻 Local	Nitro	0.469s (-22.4% 🟢)	1.004s (-1.7%)	0.535s	60	2.89x

workflow with 25 concurrent data payload steps (10KB)

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Express	0.286s (-43.9% 🟢)	1.006s (~)	0.720s	90	1.00x
🐘 Postgres	Nitro	0.295s (-40.5% 🟢)	1.006s (~)	0.711s	90	1.03x
💻 Local	Nitro	2.205s (-13.1% 🟢)	2.944s (-2.2%)	0.738s	31	7.71x
💻 Local	Express	2.226s (-11.4% 🟢)	2.943s (-2.2%)	0.717s	31	7.78x

workflow with 50 concurrent data payload steps (10KB)

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Express	0.582s (-28.9% 🟢)	1.007s (-1.0%)	0.425s	120	1.00x
🐘 Postgres	Nitro	0.634s (-19.7% 🟢)	1.006s (~)	0.372s	120	1.09x
💻 Local	Nitro	10.021s (-10.5% 🟢)	10.611s (-9.0% 🟢)	0.590s	12	17.21x
💻 Local	Express	10.042s (-10.3% 🟢)	10.694s (-10.4% 🟢)	0.652s	12	17.25x

Stream Benchmarks(includes TTFB metrics)

workflow with stream

💻 Local Development

World	Framework	Workflow Time	TTFB	Slurp	Wall Time	Overhead	Samples	vs Fastest
💻 Local	🥇 Express	1.132s (+468.5% 🔺)	2.005s (+99.6% 🔺)	0.011s (-13.2% 🟢)	2.018s (+98.2% 🔺)	0.886s	10	1.00x
💻 Local	Nitro	1.133s (+430.4% 🔺)	2.005s (+99.6% 🔺)	0.011s (-15.2% 🟢)	2.018s (+98.1% 🔺)	0.885s	10	1.00x
🐘 Postgres	Nitro	1.135s (+453.6% 🔺)	2.001s (+100.2% 🔺)	0.001s (-20.0% 🟢)	2.010s (+98.7% 🔺)	0.875s	10	1.00x
🐘 Postgres	Express	1.135s (+453.5% 🔺)	1.999s (+100.2% 🔺)	0.001s (-25.0% 🟢)	2.010s (+98.7% 🔺)	0.875s	10	1.00x

stream pipeline with 5 transform steps (1MB)

💻 Local Development

World	Framework	Workflow Time	TTFB	Slurp	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Express	1.492s (+136.8% 🔺)	2.002s (+98.9% 🔺)	0.003s (-30.4% 🟢)	2.023s (+97.8% 🔺)	0.531s	30	1.00x
🐘 Postgres	Nitro	1.519s (+143.4% 🔺)	2.004s (+99.0% 🔺)	0.004s (-2.5%)	2.027s (+98.2% 🔺)	0.507s	30	1.02x
💻 Local	Nitro	1.707s (+103.5% 🔺)	2.013s (+98.9% 🔺)	0.009s (-4.3%)	2.203s (+97.4% 🔺)	0.496s	28	1.14x
💻 Local	Express	1.923s (+154.0% 🔺)	2.011s (+95.4% 🔺)	0.009s (-8.2% 🟢)	2.422s (+132.9% 🔺)	0.499s	25	1.29x

10 parallel streams (1MB each)

💻 Local Development

World	Framework	Workflow Time	TTFB	Slurp	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Express	0.627s (-34.7% 🟢)	1.033s (-19.2% 🟢)	0.000s (-20.7% 🟢)	1.045s (-20.0% 🟢)	0.418s	58	1.00x
🐘 Postgres	Nitro	0.662s (-31.7% 🟢)	1.029s (-17.5% 🟢)	0.000s (-17.2% 🟢)	1.049s (-16.6% 🟢)	0.387s	58	1.05x
💻 Local	Express	1.365s (+11.5% 🔺)	2.015s (~)	0.000s (+40.0% 🔺)	2.018s (~)	0.652s	30	2.18x
💻 Local	Nitro	1.561s (+27.6% 🔺)	2.016s (~)	0.000s (+364.3% 🔺)	2.197s (+8.7% 🔺)	0.636s	28	2.49x

fan-out fan-in 10 streams (1MB each)

💻 Local Development

World	Framework	Workflow Time	TTFB	Slurp	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Express	1.228s (-30.7% 🟢)	1.869s (-14.2% 🟢)	0.000s (NaN%)	1.900s (-13.6% 🟢)	0.673s	32	1.00x
🐘 Postgres	Nitro	1.325s (-26.1% 🟢)	1.997s (-6.7% 🟢)	0.000s (+86.7% 🔺)	2.019s (-7.2% 🟢)	0.694s	30	1.08x
💻 Local	Express	3.158s (-8.9% 🟢)	4.028s (~)	0.001s (-33.3% 🟢)	4.030s (~)	0.873s	15	2.57x
💻 Local	Nitro	3.187s (-5.9% 🟢)	4.029s (~)	0.000s (-75.0% 🟢)	4.031s (~)	0.844s	15	2.60x

Summary

Fastest Framework by World

Winner determined by most benchmark wins

World	🥇 Fastest Framework	Wins
💻 Local	Express	12/21
🐘 Postgres	Express	17/21

Fastest World by Framework

Winner determined by most benchmark wins

Framework	🥇 Fastest World	Wins
Express	🐘 Postgres	16/21
Nitro	🐘 Postgres	17/21

Column Definitions

• Workflow Time: Runtime reported by workflow (completedAt - createdAt) - primary metric
• TTFB: Time to First Byte - time from workflow start until first stream byte received (stream benchmarks only)
• Slurp: Time from first byte to complete stream consumption (stream benchmarks only)
• Wall Time: Total testbench time (trigger workflow + poll for result)
• Overhead: Testbench overhead (Wall Time - Workflow Time)
• Samples: Number of benchmark iterations run
• vs Fastest: How much slower compared to the fastest configuration for this benchmark

Worlds:

• 💻 Local: In-memory filesystem world (local development)
• 🐘 Postgres: PostgreSQL database world (local development)
• ▲ Vercel: Vercel production/preview deployment
• 🌐 Turso: Community world (local development)
• 🌐 MongoDB: Community world (local development)
• 🌐 Redis: Community world (local development)
• 🌐 Jazz: Community world (local development)
• 🌐 Redis: Community world (local development)
• 🌐 Redis + BullMQ: Community world (local development)
• 🌐 Cloudflare: Community world (local development)
• 🌐 MySQL: Community world (local development)
• 🌐 Azure: Community world (local development)
• 🌐 NATS JetStream: Community world (local development)
• 🌐 Upstash: Community world (local development)

---

📋 View full workflow run

---

❌ Some benchmark jobs failed:

• Local: success
• Postgres: success
• Vercel: failure

Check the workflow run for details.

[github-actions]

🧪 E2E Test Results

❌ Some tests failed

Summary

	Passed	Failed	Skipped	Total
❌ ▲ Vercel Production	1196	4	219	1419
✅ 💻 Local Development	1587	0	219	1806
✅ 📦 Local Production	1458	0	219	1677
❌ 🐘 Local Postgres	1585	2	219	1806
✅ 🪟 Windows	129	0	0	129
✅ 📋 Other	727	0	176	903
Total	6682	6	1052	7740

❌ Failed Tests

▲ Vercel Production (4 failed)

astro (1 failed):

• AbortController abortFromStepWorkflow: step abort cancels an in-flight sibling step

hono (2 failed):

• outputStreamWorkflow no startIndex (reads all chunks)
• stepFunctionPassingWorkflow - step function references can be passed as arguments (without closure vars) | wrun_01KRYMPWTNKF0CE4BP27X7R946 | 🔍 observability

nextjs-webpack (1 failed):

• parallelSleepWorkflow | wrun_01KRYME5H0F7D8EH2AAF56T0Z2 | 🔍 observability

🐘 Local Postgres (2 failed)

nextjs-webpack-stable-lazy-discovery-disabled (1 failed):

• DurableAgent e2e core basic text response

nextjs-webpack-stable-lazy-discovery-enabled (1 failed):

• addTenWorkflow | wrun_01KRYMBT8QHFK9EZ9FDY03D9G0

Details by Category

❌ ▲ Vercel Production

App	Passed	Failed	Skipped
❌ astro	102	1	26
✅ example	103	0	26
✅ express	103	0	26
✅ fastify	103	0	26
❌ hono	101	2	26
✅ nextjs-turbopack	127	0	2
❌ nextjs-webpack	126	1	2
✅ nitro	103	0	26
✅ nuxt	103	0	26
✅ sveltekit	122	0	7
✅ vite	103	0	26

✅ 💻 Local Development

App	Passed	Failed	Skipped
✅ astro-stable	104	0	25
✅ express-stable	104	0	25
✅ fastify-stable	104	0	25
✅ hono-stable	104	0	25
✅ nextjs-turbopack-canary	110	0	19
✅ nextjs-turbopack-stable-lazy-discovery-disabled	129	0	0
✅ nextjs-turbopack-stable-lazy-discovery-enabled	129	0	0
✅ nextjs-webpack-canary	110	0	19
✅ nextjs-webpack-stable-lazy-discovery-disabled	129	0	0
✅ nextjs-webpack-stable-lazy-discovery-enabled	129	0	0
✅ nitro-stable	104	0	25
✅ nuxt-stable	104	0	25
✅ sveltekit-stable	123	0	6
✅ vite-stable	104	0	25

✅ 📦 Local Production

App	Passed	Failed	Skipped
✅ astro-stable	104	0	25
✅ express-stable	104	0	25
✅ fastify-stable	104	0	25
✅ hono-stable	104	0	25
✅ nextjs-turbopack-canary	110	0	19
✅ nextjs-turbopack-stable-lazy-discovery-disabled	129	0	0
✅ nextjs-webpack-canary	110	0	19
✅ nextjs-webpack-stable-lazy-discovery-disabled	129	0	0
✅ nextjs-webpack-stable-lazy-discovery-enabled	129	0	0
✅ nitro-stable	104	0	25
✅ nuxt-stable	104	0	25
✅ sveltekit-stable	123	0	6
✅ vite-stable	104	0	25

❌ 🐘 Local Postgres

App	Passed	Failed	Skipped
✅ astro-stable	104	0	25
✅ express-stable	104	0	25
✅ fastify-stable	104	0	25
✅ hono-stable	104	0	25
✅ nextjs-turbopack-canary	110	0	19
✅ nextjs-turbopack-stable-lazy-discovery-disabled	129	0	0
✅ nextjs-turbopack-stable-lazy-discovery-enabled	129	0	0
✅ nextjs-webpack-canary	110	0	19
❌ nextjs-webpack-stable-lazy-discovery-disabled	128	1	0
❌ nextjs-webpack-stable-lazy-discovery-enabled	128	1	0
✅ nitro-stable	104	0	25
✅ nuxt-stable	104	0	25
✅ sveltekit-stable	123	0	6
✅ vite-stable	104	0	25

✅ 🪟 Windows

App	Passed	Failed	Skipped
✅ nextjs-turbopack	129	0	0

✅ 📋 Other

App	Passed	Failed	Skipped
✅ e2e-local-dev-nest-stable	104	0	25
✅ e2e-local-dev-tanstack-start-	104	0	25
✅ e2e-local-postgres-nest-stable	104	0	25
✅ e2e-local-postgres-tanstack-start-	104	0	25
✅ e2e-local-prod-nest-stable	104	0	25
✅ e2e-local-prod-tanstack-start-	104	0	25
✅ e2e-vercel-prod-tanstack-start	103	0	26

---

📋 View full workflow run

---

❌ Some E2E test jobs failed:

• Vercel Prod: failure
• Local Dev: success
• Local Prod: failure
• Local Postgres: failure
• Windows: success

Check the workflow run for details.

[Copilot]

Pull request overview

Adds a new "Versioning" Foundations page (v4 and v5) explaining default run pinning to deployments, recovery via cancel + rerun, opting into newer code with deploymentId: "latest", and carrying serializable state/streams across child-run continuations. Existing references to deploymentId: "latest" and versioning behavior are linked back to the new page. Cookbook pages also pick up callouts about long-lived / multi-turn patterns and, on v5 only, are updated to reflect that start() can now be called directly inside "use workflow" functions (via #1491), while v4 retains the step-wrapped pattern.

Changes:

• New foundations/versioning.mdx for both v4 and v5, plus meta.json + index.mdx + linkbacks from vercel-world, start.mdx, and code-transform.mdx.
• v5 cookbook + migration guides updated to drop "use step" wrappers around start() in favor of calling start() directly from the workflow function; v4 keeps the step-wrapped pattern.
• Added "Versioning" callouts to scheduling, durable-agent, ai-sdk, chat-sdk, and sandbox cookbooks; fixed a few small wording/grammar issues in code-transform.mdx.

Reviewed changes

Copilot reviewed 32 out of 32 changed files in this pull request and generated no comments.

Show a summary per file

File	Description
docs/content/docs/v5/foundations/versioning.mdx	New Foundations page explaining run pinning, recovery, and explicit upgrade boundaries.
docs/content/docs/v4/foundations/versioning.mdx	v4 counterpart; uses step-wrapped continueDigest/continueExportOnLatest helpers.
docs/content/docs/v{4,5}/foundations/{meta.json,index.mdx}	Surface the new Versioning page in nav and the foundations index.
docs/content/docs/v{4,5}/deploying/world/vercel-world.mdx	Link the deployment-pinning section to the new Versioning page.
docs/content/docs/v{4,5}/how-it-works/code-transform.mdx	Link "atomic versioning" reference; fix "function assume" / "from being upgraded" wording.
docs/content/docs/v{4,5}/api-reference/workflow-api/start.mdx	Soften "Vercel-specific" to "Vercelism", mention possible version rename, link Versioning. v5 also notes native workflow-context start().
docs/content/docs/v5/foundations/starting-workflows.mdx	Note v5 can call start() from workflows; link Composition + Versioning.
docs/content/docs/v5/cookbook/common-patterns/workflow-composition.mdx	Replace step-wrapped child spawn with direct start() call.
docs/content/docs/v5/cookbook/advanced/child-workflows.mdx	Inline batched start() calls; rename spawnReportChunk → startReportChunk; update tips.
docs/content/docs/v{4,5}/cookbook/common-patterns/scheduling.mdx	Callout linking to Versioning for long-lived schedules.
docs/content/docs/v{4,5}/cookbook/agent-patterns/durable-agent.mdx	Callout linking to Versioning for multi-turn agents.
docs/content/docs/v{4,5}/cookbook/integrations/{ai-sdk,chat-sdk,sandbox}.mdx	Callouts linking to Versioning for long-lived sessions.
docs/content/docs/v4/cookbook/common-patterns/workflow-composition.mdx	Update deploymentId: "latest" callout wording + link to Versioning.
docs/content/docs/v4/cookbook/advanced/child-workflows.mdx	Link Versioning from the deploymentId: "latest" tip.
docs/content/docs/v4/changelog/eager-processing.mdx	Reword to "step-wrapped start() helper" to match v4 phrasing.
docs/content/docs/v5/migration-guides/{migrating-from-trigger-dev,migrating-from-temporal,migrating-from-inngest,migrating-from-aws-step-functions}.mdx	Update to v5 native workflow-context start(); only collect step retains "use step".

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[TooTallNate]

Comment-only review

The new versioning foundations page is a useful addition — the four sections (default pinning → recovery → self-upgrade via start({ deploymentId: "latest" }) → carrying serializable context forward) cover the model clearly, and the v4-vs-v5 differentiation (step-wrapped start() vs direct start() from workflow context, gated on whether PR #1491 has landed on the target branch) is correctly handled. The cross-links from cookbook patterns to the new foundations page tie everything together nicely.

Three observations — see inline.

Inline comments

1. Wording: this PR introduces "Vercelism" as a new term and uses it 6 times across the diff, including replacing existing "Vercel-specific feature" wording in two <Callout>s. "Vercelism" reads as informal slang and isn't part of the established docs vocabulary. The existing tone uses "Vercel-specific", "Vercel deployments", "Vercel World", "non-Vercel" — all precise. Suggested restoring the prior phrasing. This is the only thing I'd consider blocking.
2. Broken link in v4: versioning.mdx references /docs/foundations/cancellation in its related frontmatter, but that page is v5-only. Easy fix.
3. Missing changeset: convention for docs-only PRs is an empty ---\n--- file.

A few small things I noticed

• The pre-emptively check for a conflict with getHookByToken() example (line 96 of versioning.mdx) — nice that you noted the race window in a later doc (#2011 callout), but this page doesn't have that warning. Worth a cross-link or <Callout type="warn"> here too, since the recovery pattern uses both start() and deploymentId: "latest" together.
• The _idempotency / using syntax is great. Worth confirming the docs site's syntax highlighter renders it — I noticed #1988 recently fixed a homepage link lint regression, and the using syntax is TypeScript 5.2+ specific.
• dailyDigest example (line 137-152): the state parameter is DigestState but the workflow doesn't actually use state.userId after the first sendDigest call — pattern is fine, just noting if you wanted to make the example illustrate state evolution more vividly, having sendDigest read state.lastSentAt (e.g. to avoid sending duplicates) would make the carry-forward intent clearer.

Overall the page reads well — once the Vercelism wording is restored and the v4 link is fixed, this should be good to go.

[TooTallNate]

All three review comments addressed in 7eff96c — Vercelism wording reverted, v4 broken cancellation link fixed, empty changeset added. LGTM, approving.

[github-actions]

Backport PR opened against stable: #2014. Merge conflicts were resolved by AI — please review carefully.