This repository contains a Human-Agent Experience (HAX) Interface, designed to enhance collaboration between humans and AI agents.
The HAX Interface provides tools and guidelines for creating consistent, trusted, and scalable collaboration experiences. It focuses on enabling agents to control UI elements, proactively present information, and seamlessly involve humans in decision-making processes.
The HAX Interface architecture is designed to facilitate seamless human-agent collaboration. Key components include:
- HAX SDK: A schema-driven interface layer where agents send structured data to frontend components using a consistent format.
- Reusable UI Components: Pre-built components for common interaction patterns (e.g., confidence indicators, decision logs, action asks).
- HAX Playbooks: Pre-defined behavioral patterns for agents to follow in specific scenarios.
- Looped-In-Human (LIH) Mechanisms: Tools for agents to involve humans in decision-making when needed.
The HAX SDK includes a CLI to init setup, add components, and more.
To get started with the CLI, see the CLI README for installation and usage instructions.
See the open issues for a list of proposed features (and known issues). (Replace with the actual repo issues link)
Contributions are what make the open source community such an amazing place to learn, inspire, and create. Any contributions you make are greatly appreciated. For detailed contributing guidelines, please see CONTRIBUTING.md
Distributed under the Apache-2.0 License. See LICENSE for more
information.
For more information about our various agents, please visit the Outshift Open page.
`