Alpha Release Notice: This SDK is currently in beta status (v0.1.0-beta). The API is still evolving and may change in future releases as we gather feedback and improve functionality.

Visit our homepage hud.so →

​
Overview

HUD provides an interface for:

  • Creating and running evaluation environments for browser and computer use agents

  • Recording agent interactions through detailed trajectories

  • Managing evaluation jobs across multiple tasks

  • Supporting different agent adapters

  • Providing telemetry to track agent performance

​
Key Concepts

  • Environment - A running instance where an agent can interact with code execution capabilities and built-in telemetry

  • Task - Configuration for reproducibly creating an evaluation environment with a defined problem statement

  • Trajectory - Record of agent actions, observations, and environment states during a runthrough

  • Job - Collection of related trajectories for evaluating agent performance across multiple tasks

​
Get Started

Follow our Installation Guide to get set up, then check out the Quickstart Guide to run your first example.