Vercel has released updates to its AI SDK, including version @ai-sdk/react@4.0.47, @ai-sdk/tui@1.0.45, @ai-sdk/workflow@1.0.44, and @ai-sdk/rsc@3.0.44. These releases address a telemetry issue and introduce new capabilities for Amazon Bedrock integration.
Key Points
- The ai@7.0.44 package includes a fix for telemetry attribution for language model calls.
- This fix addresses cases where language model calls resolve to a different response model.
- The @ai-sdk/amazon-bedrock@5.0.38 package now supports video inputs in Converse messages.
- Several other @ai-sdk packages, including @ai-sdk/angular@3.0.44 and @ai-sdk/code-mode@1.0.1, have updated dependencies on ai@7.0.44.
- The @ai-sdk/harness family of packages, such as @ai-sdk/harness-claude-code@1.0.51, also updated their dependencies.
Context
According to the Vercel AI SDK Changelog, these updates were released via the Changesets release GitHub action. The process involves merging a pull request to publish packages to npm automatically. The changelog indicates that if not ready for release, the pull request will update as more changesets are added to the main branch.
Why It Matters
The telemetry fix ensures more accurate data collection for language model usage, which can be important for developers monitoring their applications. The addition of video input support for Amazon Bedrock's Converse messages expands the multimodal capabilities available to builders using the Vercel AI SDK, allowing for more complex interactions.
What To Do
- Review the updated @ai-sdk/amazon-bedrock@5.0.38 documentation for details on implementing video inputs in Converse messages.
- Check the ai@7.0.44 changelog to understand the specifics of the telemetry attribution fix.
- Update your project dependencies to the latest @ai-sdk package versions to incorporate these changes.
- Test existing language model call implementations to ensure telemetry attribution functions as expected with the new fix.