​
Judgeval SDK (TypeScript)

  • Published TypeScript SDK on npm
  • Added agent cookbooks for TypeScript integration
  • Integrated OpenTelemetry with Vercel AI SDK

​
Infrastructure & Self-Hosting

  • Added clustering support for experiment debugging with automated topic failure analysis
  • Added S3 bucket upload support for trace data via SDK
  • Released command-line tool for self-hosted environment setup

​
Experimentation & Analysis

  • Added load diff feature for large-scale experiment comparisons
  • Enabled single and last_k trace analysis for MCP server

​
UI Improvements

  • Added quickstart popup with tracing and experiment cookbooks

​
LLM Features

  • Added async LLM client tracing support