Sign in
← AI PulseSep 16, 2026

Deep · news · Multi-source brief

Vercel AI SDK Adds TypeSafe Provider for Experimental Evaluation

The Vercel AI SDK has introduced a TypeSafe provider for experimental_evaluate, enabling native Choice, Score, and Boolean questions, structured JSON rubrics, and workflow serialization.

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/langchain@3.0.103 · vercel/ai
Save

The Vercel AI SDK has added a TypeSafe provider for experimental_evaluate, according to recent changelog entries for @ai-sdk/langchain@3.0.103, @ai-sdk/llamaindex@3.0.103, @ai-sdk/klingai@4.0.44, and @ai-sdk/mcp@2.0.51. This new provider supports native Choice, Score, and Boolean questions within a single request, alongside structured JSON rubrics.

Key Points

  • The TypeSafe provider for experimental_evaluate is now available in the Vercel AI SDK.
  • It supports native Choice, Score, and Boolean questions in a single request.
  • Structured JSON rubrics are included for evaluation.
  • The provider also handles probability distributions and confidence metadata.
  • Workflow serialization is a supported feature.
  • Mid-conversation tool discovery and updates for code-mode have been added to the ai package.
  • Node.js downloads are protected by DNS validation and connection pinning, even when frameworks wrap global fetch.

Context

According to the Vercel AI SDK changelogs, the TypeSafe provider for experimental_evaluate was introduced as a major change. This enhancement allows for a more structured approach to evaluating AI model outputs by supporting specific question types and JSON-based rubrics. The updates also include improvements to the core ai package, such as tool discovery during conversations and enhanced security for Node.js downloads.

Why It Matters

This update provides builders with more robust and structured tools for evaluating AI model performance, particularly for tasks requiring specific answer formats or confidence levels. The ability to define structured JSON rubrics and serialize workflows can streamline the development and testing of AI applications.

What To Do

  • Review the Vercel AI SDK documentation for details on the experimental_evaluate function and the TypeSafe provider.
  • Experiment with defining structured JSON rubrics for model evaluation tasks.
  • Investigate the new capabilities for mid-conversation tool discovery in code-mode.
  • Note the enhanced security measures for Node.js downloads related to DNS validation and connection pinning.