# ag-ui-protocol/ag-ui

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

Repository: https://github.com/ag-ui-protocol/ag-ui
Canonical: https://ross.abutalabs.com/products/ag-ui
Homepage: https://ag-ui.com
Language: Python
License: MIT
License Family: permissive
Topics: ag-ui-protocol, agent-frontend, agent-ui, agentic-workflow, ai-agents
Last push: 2026-08-26T23:04:27+00:00

## Health v2 (maintenance only)
Score: 86/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 99, release rhythm 98, longevity 34
- inputs: {"age_days": 483, "days_push": 7, "days_rel": 13, "gap_med": 2, "n_releases_24m": 44}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 15562, forks 1405 (observed 2026-08-28T04:11:11.896035+00:00)

## What it is
AG-UI is an open, lightweight, event-based protocol that standardizes how AI agent backends connect to user-facing frontend applications. It defines ~16 standard event types, a flexible middleware layer supporting any event transport (SSE, WebSockets, webhooks), and ships SDKs plus a reference HTTP implementation for building agentic user experiences.

## Use cases
- connect ai agents to frontend applications
- stream agent events to a user interface
- build a chat interface backed by an agent backend
- standardize agent state and ui intent communication
- integrate generative ui components with agents
- build a terminal cli agent client
- wire an agentic backend into a web app without ad-hoc code

## When to choose
- you need a standardized, transport-agnostic event protocol between agents and UIs
- you want interoperability across agent frameworks and frontend clients
- you are building real-time, interactive agentic features in web or chat apps
- you want to support generative UI specs like A2UI, MCP-UI, or Open-JSON-UI

## When to avoid
- you only need agent-to-tool connections (use MCP instead)
- you need agent-to-agent coordination (use A2A instead)
- you need a fully rendered UI out of the box rather than a protocol layer
- your project has no agentic or LLM component

## Facets
- artifact type: framework
- maturity: active
- function: agent-framework, middleware, websocket, http-server, sdk, chat-interface
- domain: large-language-models, web-development, frontend, apis, developer-tools
- platform: python, cross-platform, cli
- tags: agent-user-interaction, event-based-protocol, agentic-frontend, generative-ui, protocol-standard, real-time-events, copilotkit, ai-agents, nodejs, web-server

## Member repositories
- ag-ui-protocol/ag-ui (main) score 86

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:11:11.896035+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-29T17:05:39.286064+00:00, confidence not recorded.
  - readme: https://github.com/ag-ui-protocol/ag-ui (fetched 2026-08-28T04:11:11.896035+00:00, sha bde65a5af79a)
  - homepage: https://ag-ui.com (fetched 2026-08-29T08:03:49.198385+00:00, sha 2fdabd43b43c)
  - site_page: https://docs.ag-ui.com/concepts/generative-ui-specs (fetched 2026-08-29T08:03:49.207903+00:00, sha 1fce352e49de)
  - site_page: https://docs.ag-ui.com/quickstart/clients (fetched 2026-08-29T08:03:49.209802+00:00, sha aadda43951fc)
- Data as of 2026-08-30T08:39:29.467469+00:00.