AI Devtools Radar

Braintrust

LLM observabilitywww.braintrust.dev6 published changes

2 sources under watch

Change timeline

feature

Support for Zod v4 schemas in eval and function parameters

Developers can now use Zod v4 for parameter validation in evaluations and functions, improving type safety.

Evidence
+feat: Support zod v4 schemas for eval parameters and function parameters

github.com · also seen at 1 more source

feature

Add instrumentation for @typesafe-ai/sdk

Developers using the TypeSafe AI SDK can now have their SDK calls instrumented and tracked through Braintrust.

Evidence
+feat: Add instrumentation for <code>@typesafe-ai/sdk</code>

github.com

feature

Add instrumentation for synchronous ElevenLabs SDK APIs

Developers using the ElevenLabs SDK will have their synchronous API calls automatically instrumented and tracked by Braintrust.

Evidence
+feat: Add instrumentation for synchronous elevenlabs SDK APIs

github.com

pricinghigh

Monthly included credits reduced from $249 to $100

Paying users on Braintrust's main plan will receive fewer included credits per month, requiring either more frequent purchases or reduced evaluation volume.

Evidence
−$249 credits / month includedthen token rates
+$100 credits / month includedthen token rates

www.braintrust.dev

deprecationhigh

@braintrust/vercel-ai-sdk package is now deprecated with v0.0.7 being the final release

Users of @braintrust/vercel-ai-sdk must migrate to using wrapAISDK from the braintrust package or automatic instrumentation via runtime hooks. No further updates will be provided for this package.

Evidence
+@braintrust/vercel-ai-sdk is now deprecated. This release marks the last release for this package.

github.com

feature

Add experimental batch evals API to Braintrust SDK

Developers can now use an experimental batch evaluations API for running batch evaluations.

Evidence
+feat: Add experimental batch evals API (#2297)

github.com · also seen at 1 more source