Sign up to save tools and stay up to date with the latest in AI
bg
bg
1

GitHub - getevalkit/evalkit

May 19, 2025 - github.com
EvalKit is an open-source dashboard designed to help domain experts enhance AI agents. It offers features such as tracing interactions and tasks to improve the performance of AI systems. The Quickstart guide provides instructions for installing and setting up the core library using either `uv` or `pip`, and demonstrates how to instrument agent code with EvalKit for tracing purposes.

To get started, users need to configure the server backend, which runs on Vertex AI. This involves setting environment variables, creating a service account JSON file, and running the backend on a local server. The frontend setup requires installing dependencies and running the development server. An example is provided to demonstrate the core library's features, which can be executed using the `uv` command.

Key takeaways:

  • EvalKit is an open-source dashboard designed to enhance AI Agents through domain expert input.
  • The Quickstart guide includes installing the core library and instrumenting agent code with EvalKit.
  • Server setup requires configuring Vertex AI and setting environment variables, with the backend running on http://localhost:8000.
  • The frontend setup involves running npm commands, with the frontend accessible at http://localhost:5173.
View Full Article

Comments (0)

Be the first to comment!