Vercel has updated its AI SDK, addressing telemetry attribution for language model calls and introducing new capabilities for Amazon Bedrock. These changes are part of a series of patch releases across various SDK components, as detailed in the Vercel AI SDK Changelog.
Key Points
- The AI SDK (
ai@7.0.44) received a patch to fix telemetry attribution for language model calls that resolve to a different response model. - The Amazon Bedrock SDK (
@ai-sdk/amazon-bedrock@5.0.38) now supports video inputs in Converse messages. - Several harness packages, including
@ai-sdk/harness-claude-code@1.0.51and@ai-sdk/harness-codex@1.0.52, were updated. - Other packages, such as
@ai-sdk/angular@3.0.44,@ai-sdk/code-mode@1.0.1, and@ai-sdk/langchain@3.0.44, also received dependency updates.
Context
According to the Vercel AI SDK Changelog, these updates are part of routine releases managed by a GitHub action. The process involves merging a pull request to publish packages to npm automatically. The changelog indicates that if a release is not ready, the pull request will update whenever more changesets are added to the main branch.
Why It Matters
These updates enhance the Vercel AI SDK's functionality and reliability. The telemetry fix ensures more accurate data collection for model performance, while video input support for Amazon Bedrock expands the types of multimodal interactions developers can build.
What To Do
- Developers using the Vercel AI SDK should update their
aipackage to version 7.0.44 to benefit from the telemetry fix. - Those integrating with Amazon Bedrock should update
@ai-sdk/amazon-bedrockto version 5.0.38 to enable video input in Converse messages. - Review the specific changelog entries for any other
@ai-sdkpackages in use to understand dependency updates. - Watch the Vercel AI SDK GitHub repository for future updates and release announcements.
