Demos, ideas, and reusable assets for building with A2UI (Agent-to-User Interface), the open protocol that lets AI agents describe rich, interactive UIs that client apps render natively.
A2UI and A2UI Labs are vendor-, model-, and framework-agnostic open-source projects—not a Google- or Gemini-only product. The protocol and reference applications in this repository are designed to work across:
- Any UI framework: Reference apps and renderers span Flutter, React, Web Components, and other native or web client stacks.
- Any AI model or agent backend: Demos and assets can be paired with any LLM provider, local model, or agent orchestration framework. No Google account, Google Cloud service, or Google API is required to use A2UI or run these patterns.
- Demos: end-to-end reference applications (such as travel and personal-finance showcases) pairing client UIs with pluggable agent backends.
- Ideas: experiments and prototypes that explore new A2UI patterns.
- Assets: reusable components, templates, and snippets to use in your own A2UI projects.
Each demo, idea, and asset lives in its own directory with a README that explains how to run or use it, including how to configure your preferred model or backend.
See CONTRIBUTING.md for how to contribute.
Apache 2.0. See LICENSE.
This project is intended for demonstration purposes only. It is not intended for use in a production environment.
This project is not intended for active development. Issues and contributions are handled on a best effort basis.
Eligibility for the Google Open Source Software Vulnerability Rewards Program is determined by the Google Open Source Software Vulnerability Reward Program Rules.
While A2UI Labs is model- and vendor-agnostic, some individual demos may optionally use the Gemini API or other external services as an example backend. Any optional use of the Gemini API is subject to the Gemini API Additional Terms of Service, and each demo documents the Terms of Service for any optional external services it calls in its own README.