ag-ui-protocol/ag-ui

AG-UI: the Agent-User Interaction Protocol. Bring Agents into Frontend Applications.

What it solves

AG-UI provides a standardized, lightweight protocol for connecting AI agents to user-facing applications. It eliminates the need for custom, fragmented integration logic when bringing agents into a UI, allowing for seamless interaction between agent backends and frontends.

How it works

AG-UI is an event-based protocol. Agent backends emit events that match one of approximately 16 standard event types, and accept a few simple compatible inputs. It includes a flexible middleware layer that allows it to work across various event transports (such as WebSockets, SSE, or webhooks) and supports loose event format matching to ensure interoperability between different agents and applications.

Who it’s for

Developers building AI agents and application developers who want to integrate these agents into user interfaces with features like real-time streaming, generative UI, and human-in-the-loop collaboration.

Highlights

  • Broad Framework Support: Integrates with major frameworks like LangChain, CrewAI, LlamaIndex, and Pydantic AI.
  • Multi-Language SDKs: Provides SDKs for Kotlin, Golang, Dart, Java, Rust, Ruby, C++, and .NET.
  • ** uma-directional State Sync**: Supports bi-directional state synchronization and real-time context enrichment.
  • Generative UI: Enables structured messages and generative UI components within the agent interaction.
  • Transport Agnostic: Works with any event transport mechanism.

Related

  • Project
    a2aproject/A2A

    An open protocol that enables communication and interoperability between AI agents built on different frameworks and servers, allowing them to collaborate without exposing internal logic.

  • Project
    SolaceLabs/solace-agent-mesh

    An open-source framework for building event-driven multi-agent AI systems using the Solace Platform for scalable and reliable agent communication.

  • Project
    machinepulse-ai/world2agent

    An open protocol and framework that standardizes how AI agents perceive the real world by providing structured, real-time data via a system of interchangeable sensors.

  • Project
    ag2ai/ag2

    An open-source AgentOS framework for building AI agents and orchestrating their cooperation through a hub-and-channel network system.

  • Project
    Haohao-end/openagent

    A full-stack AI agent platform for building, orchestrating, and publishing AI applications with visual workflows, RAG integration, and OpenAPI delivery.