Sign in
← AI PulseSep 16, 2026

Wire · news · Multi-source brief

Vercel AI SDK Adds TypeSafe Provider for Experimental Evaluation, Enhances Tool Discovery

The Vercel AI SDK has introduced a TypeSafe provider for experimental_evaluate, enabling native Choice, Score, and Boolean questions in a single request, alongside updates for mid-conversation tool discovery in code-mode.

By Illumora Editorial

Source · Sep 16, 2026, 6:26 PM · On Illumora · Sep 16, 2026, 6:39 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/mcp@2.0.51 · vercel/ai
Save

The Vercel AI SDK has released updates across several packages, including @ai-sdk/mcp@2.0.51, @ai-sdk/huggingface@2.0.50, @ai-sdk/hume@3.0.42, and @ai-sdk/klingai@4.0.44. A significant update is the addition of a TypeSafe provider for experimental_evaluate in @ai-sdk/typesafe-ai@3.0.0, which supports various question types and evaluation features. Additionally, the ai@7.0.103 package includes new capabilities for tool discovery.

Key Points

  • The @ai-sdk/typesafe-ai@3.0.0 package now includes a TypeSafe provider for experimental_evaluate.
  • This provider supports native Choice, Score, and Boolean questions within a single request.
  • Structured JSON rubrics, probability distributions, and confidence metadata are also supported by the new TypeSafe provider.
  • Workflow serialization is a feature of the experimental_evaluate TypeSafe provider.
  • The ai@7.0.103 package adds mid-conversation tool discovery and updates for code-mode.
  • Node.js downloads maintain DNS validation and connection pinning even when frameworks wrap global fetch.
  • The rawInput in output-error UI message parts has been deprecated.

Context

According to the Vercel AI SDK changelog, these updates aim to enhance the evaluation capabilities and developer experience within the SDK. The experimental_evaluate feature, now bolstered by the TypeSafe provider, offers more structured and comprehensive ways to assess model outputs. The improvements to tool discovery in code-mode reflect an ongoing effort to refine how AI models interact with and utilize external tools during conversations.

Why It Matters

These updates provide developers with more robust tools for evaluating AI model performance and integrating dynamic tool use. The enhanced experimental_evaluate functionality can lead to more precise and detailed assessments of model responses, while improved tool discovery can streamline the development of more interactive and capable AI applications.

What To Do

  • Note the availability of the TypeSafe provider for experimental_evaluate in @ai-sdk/typesafe-ai@3.0.0.
  • Review the new capabilities for native Choice, Score, and Boolean questions in evaluation workflows.
  • Consider the implications of mid-conversation tool discovery for applications using code-mode.
  • Observe the deprecation of rawInput in output-error UI message parts for future updates.