Sign in
← AI PulseSep 11, 2026

Wire · news · Multi-source brief

Vercel AI SDK Updates Workflow, XAI, and Workflow Harness Packages

Vercel has released updates for its AI SDK, including version 2.0.29 of @ai-sdk/workflow, 4.0.58 of @ai-sdk/xai, and 1.0.108 of @ai-sdk/workflow-harness, all released on September 11.

By Illumora Editorial

Source · Sep 11, 2026, 6:59 PM · On Illumora · Sep 11, 2026, 7:03 PM

Media from the primary source — shown here so you can stay on Illumora.

Synthesized from multiple allowlisted primaries on the same event. Lanes →

Brief drafted by Illumora’s editorial model from the linked primary source. Ops desk reviews flagged pieces. How we write →

Read the source →Vercel AI SDK Changelog — Release @ai-sdk/workflow@2.0.29 · vercel/ai
Save

Vercel released several updates to its AI SDK on September 11, including @ai-sdk/workflow@2.0.29, @ai-sdk/xai@4.0.58, and @ai-sdk/workflow-harness@1.0.108. These updates introduce fixes and new features across the respective packages, addressing tool result preservation, batch request handling, and workflow resumability.

Key Points

  • @ai-sdk/workflow@2.0.29 fixes an issue where provider-executed tool results were not preserved across terminal, deferred, and paused responses.
  • @ai-sdk/xai@4.0.58 adds support for image generation requests within batches.
  • @ai-sdk/xai@4.0.58 also includes a fix to reject unsupported batch request types.
  • A fix in @ai-sdk/xai@4.0.58 for DeepSeek models preserves reasoning streams even with empty tool-call deltas.
  • @ai-sdk/workflow-harness@1.0.108 ensures resumability is maintained when harness runs encounter failures.
  • @ai-sdk/workflow-harness@1.0.108 now automatically persists validated HarnessAgent output within workflow results.

Context

According to the Vercel AI SDK changelog, these updates are part of ongoing maintenance and feature development for the SDK. The @ai-sdk/workflow package focuses on managing the state and execution of AI workflows, while @ai-sdk/xai appears to handle extended AI functionalities, including batch processing and specific model integrations like DeepSeek. The @ai-sdk/workflow-harness package provides tools for testing and validating AI workflow components.

Why It Matters

These updates enhance the robustness and capabilities of the Vercel AI SDK. Developers using the SDK can expect more reliable tool execution within workflows, expanded support for batch processing of image generation tasks, and improved resilience in workflow testing environments. The specific fix for DeepSeek reasoning streams indicates attention to model-specific behaviors.

What To Do

  • Review the changelog for @ai-sdk/workflow@2.0.29 to understand the implications of preserved tool results.
  • Test batch image generation requests using @ai-sdk/xai@4.0.58 if your application utilizes this feature.
  • Note the fix for DeepSeek reasoning streams in @ai-sdk/xai@4.0.58 if you are integrating with DeepSeek models.
  • Examine the updates in @ai-sdk/workflow-harness@1.0.108 regarding resumability and automatic output persistence for workflow testing.