6d7e5035a4
Version Packages (beta) ( #711 )
2025-12-31 10:17:49 -08:00
307f4b0e41
Revert "Add lazy workflow and step discovery in Next.js ( #640 )" ( #712 )
2025-12-31 09:37:56 -08:00
7ff68d1753
Pass tsconfig to esbuild for support of "paths" aliases ( #705 )
2025-12-31 01:00:24 -08:00
9b1640d76e
Do not include initial attempt in step function maxRetries count ( #703 )
...
Closes #592 .
2025-12-31 00:59:50 -08:00
b49bb4eb54
Version Packages (beta) ( #707 )
2025-12-30 11:01:55 +01:00
851c0203fe
Validate cached workflows/steps and reduce build debounce ( #700 )
2025-12-30 10:36:35 +01:00
2dbe49495d
Fix race condition in streamer when multiple writes share a promise runId ( #704 )
...
ULID generation now happens synchronously before awaiting the runId, ensuring chunks maintain their call order.
2025-12-30 10:35:18 +01:00
b21d0b35cc
[docs] Add PR template for PRs not created through the quick PR feature ( #702 )
...
* [docs] Add PR template for OSS contributions
Signed-off-by: Peter Wielander <mittgfu@gmail.com >
* Add line
Signed-off-by: Peter Wielander <mittgfu@gmail.com >
* Empty
Signed-off-by: Peter Wielander <mittgfu@gmail.com >
---------
Signed-off-by: Peter Wielander <mittgfu@gmail.com >
2025-12-29 13:04:37 -06:00
4cf671f083
docs: add getWorld ( #559 )
...
* docs: add getWorld
* update
2025-12-29 12:48:21 -06:00
62cde6dd8d
Version Packages (beta) ( #696 )
2025-12-28 18:07:16 +01:00
25bfa52d02
Fix: Handle object-style finishReason for AI SDK v5/v6 compatibility ( #695 )
...
---------
Signed-off-by: Peter Wielander <mittgfu@gmail.com >
Co-authored-by: Peter Wielander <mittgfu@gmail.com >
2025-12-28 18:00:25 +01:00
80955e7212
[cli] [web] Allow opening web UI without config validation, show better UI when runs can't be found ( #684 )
2025-12-28 17:55:24 +01:00
f989613d70
[web] Add error boundaries around UI in run detail view ( #694 )
...
Signed-off-by: Peter Wielander <mittgfu@gmail.com >
2025-12-28 14:08:11 +01:00
e6b1bef5bb
Version Packages (beta) ( #693 )
2025-12-27 12:40:09 -08:00
e3f0390469
Workflows graph extractor ( #455 )
...
---------
Signed-off-by: Karthik Kalyanaraman <karthik@scale3labs.com >
2025-12-27 16:52:19 +01:00
b97b87b4fe
Add AI SDK v6 support ( #692 )
2025-12-27 16:24:54 +01:00
96a4d3a2fd
Version Packages (beta) ( #676 )
...
Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>
2025-12-24 09:51:14 -08:00
26d9769335
ai sdk parity for DurableAgent ( #668 )
...
### Key Changes
**Generation Settings**
- Support for `temperature`, `maxOutputTokens`, `topP`, `topK`, `presencePenalty`, `frequencyPenalty`, `stopSequences`, `seed` at constructor and per-stream level
**Step Control**
- Added `maxSteps` to limit LLM calls (default: unlimited for backwards compatibility)
- Added `toolChoice` (`auto`, `none`, `required`, or specific tool)
- Added `activeTools` to filter available tools per-stream call
**Callbacks**
- `onChunk` - called for each stream chunk
- `onStepFinish` - called after each step completes
- `onFinish` - called when all steps complete
- `onError` - called on errors
- `onAbort` - called when aborted via `AbortSignal`
**Structured Output**
- Added `experimental_output` with `Output.object({ schema })` for parsing structured responses
**Context & Telemetry**
- Added `experimental_context` to pass data to tool executions
- Added `experimental_telemetry` for observability
**Stream Result**
- `stream()` now returns `{ messages, steps, experimental_output }` instead of just `{ messages }`
**Other**
- `prepareStep` can now override generation settings, toolChoice, and context
- Added `experimental_repairToolCall` for handling malformed tool calls
- Enabled file, source, and tool input streaming chunks
- Exported all new types from `@workflow/ai`
2025-12-24 09:37:02 -08:00
d9f6a49397
[world-local] Create dataDir on start and add versioning forward-compat ( #681 )
2025-12-23 20:43:03 +01:00
0cf0ac3211
[util] Extract helper function to find local world dataDir across CLI/UI ( #682 )
2025-12-23 20:42:42 +01:00
8ba8b6be6b
Revert extra stdlib re-export ( #683 )
2025-12-23 12:04:47 -06:00
25b02b0bfd
[runtime] Add CORS headers and request handling to health check check response ( #624 )
2025-12-23 14:17:17 +01:00
c059cf6fcd
[cli] Only do interactive pagination when using the -i flag ( #669 )
2025-12-23 14:05:06 +01:00
4bc98ff4a1
[web-shared] Move error boundary from web to web-shared and use for sidebar detail view ( #674 )
2025-12-23 14:02:36 +01:00
ea3afce222
Add lazy workflow and step discovery in Next.js ( #640 )
2025-12-22 23:16:17 -06:00
3f0a2e76d2
Bring back agent sitemap code ( #665 )
...
Original PR #108
This was deleted in a bad merge here https://github.com/vercel/workflow/commit/c0991b4ff795687bb477e96f6b4261a2c8a5bc61#diff-4f67490b6b20a5606964421a564836c3033f0ff2b0d90a41fe8822001d217f20
Also introduces proper canonical URLs and alternates for the MD responses.
Signed-off-by: Malte Ubl <malte.ubl@gmail.com >
2025-12-22 11:17:19 -08:00
c3464bfd97
Fix race conditions in webhook response handling ( #672 )
...
Fixes flaky `webhookWorkflow` e2e test on "local" test suites
2025-12-22 10:41:08 -08:00
0143da749d
Version Packages (beta) ( #675 )
2025-12-22 16:22:58 +01:00
cbf9b46aea
[ci] Github changelog: fix packages that haven't changed being re-listed with same changes ( #664 )
2025-12-22 16:19:53 +01:00
616bc67be4
[web-shared] Fix null access in eventData ( #673 )
2025-12-22 16:18:17 +01:00
560dda3014
Version Packages (beta) ( #660 )
2025-12-20 10:58:29 +01:00
696ef450f2
[ci] Fix github changelog generation missing some PR links ( #663 )
2025-12-20 10:50:12 +01:00
f2d5997b80
[world-local] Throw an error when trying writing JSON that fails entity validation
2025-12-20 10:44:00 +01:00
88ad5c9bbf
[web] Fix hook span display if hooks were already disposed ( #658 )
2025-12-20 09:52:32 +01:00
ef22f82c9e
[web] Allow resuming hooks with payloads from the UI ( #656 )
2025-12-20 09:27:01 +01:00
bd3385ff6d
Version Packages (beta) ( #652 )
...
Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>
2025-12-19 15:41:36 -08:00
75a5060473
Handle Vercel Queue 24 hour message TTL by re-enqueueing message ( #651 )
2025-12-19 15:30:51 -08:00
eaf9aa65f3
Don't shadow globalThis ( #649 )
2025-12-18 15:37:52 -08:00
6cd1a47b31
Additional error debugging when failing to validate server response ( #647 )
2025-12-18 15:17:33 -08:00
f39683370d
fix: include next config in web package ( #646 )
...
* fix: include next config in web package
* changeset
2025-12-18 15:10:45 -08:00
ea3815c75d
docs: mention nextjs breaking change with workflow compatibility ( #645 )
2025-12-18 13:41:54 -08:00
187225fa64
Fix release target ( #644 )
2025-12-18 20:46:38 +01:00
2a3bbb79e8
Fix releasing as latest on npm, but not within changeset ( #643 )
2025-12-18 20:41:46 +01:00
b4132d0895
Version Packages (beta) ( #635 )
2025-12-18 20:07:06 +01:00
28224fff13
Compile release note summaries using custom script in CI ( #637 )
2025-12-18 19:06:05 +01:00
ea2a67e19c
perf: parallelize suspension handler for high-concurrency ( #544 )
...
The current implementation of suspension handler processes the invocation queue sequentially which is really slow. Now we process is in parallel:
- Process hooks first, then steps and waits in parallel to prevent race conditions
- Refactor runtime.ts into modular files: `suspension-handler.ts`, `step-handler.ts`, `helpers.ts`
- nit: Add otel attributes for hooks created (`workflow.hooks.created`) and waits created (`workflow.waits.created`)
- nit: Update suspension status from `pending_steps` to `workflow_suspended`
2025-12-17 18:43:09 -08:00
ab55ba2d61
fix: disable RSC dedupe request through world-vercel ( #639 )
...
* fix: add request header for world-vercel
* changeset
2025-12-17 15:43:10 -08:00
712f6f86b1
[worlds] [runtime] Change serialized stream ID gen from v4 UUIDs to ULIDs and fix stream list endpoint ( #625 )
2025-12-17 20:44:34 +01:00
4bdd3e5086
Move auth error messages into @workflow/errors package ( #638 )
2025-12-17 16:54:29 +01:00
c6f33ee9d3
[web] Show event markers for step_started events ( #636 )
2025-12-17 10:35:09 +01:00
Vercel Release Bot