---
id: ag-ui-langgraph
version: "0.0.42"
license: MIT
license_treatment: permissive
maintenance: active
---

# ag-ui-langgraph — Implementation of the AG-UI protocol for LangGraph.

License: permissive · Maintenance: active · Downloads: 801.3K/mo

## What it is and what it does

ag-ui-langgraph is a protocol adapter that connects LangGraph agent workflows to the AG-UI (Agent User Interface) protocol standard. It provides an integration layer that automatically exposes LangGraph graphs as HTTP endpoints, handling the translation between LangChain's message format and the AG-UI event model. The package handles comprehensive event streaming—including agent thinking, tool invocations, and state updates—so frontend applications can display live agent execution without building custom endpoint logic.

The package sits between your LangGraph workflow and a web client, managing protocol conversion and event delivery. It depends on langchain, langgraph, pydantic, and ag-ui protocol libraries. Installation is straightforward (pure wheel, low friction), and the API is minimal: pass your graph and a path to add_langgraph_fastapi_endpoint() and the endpoint is ready. It's designed for teams building agent UI applications that need a standard protocol bridge rather than custom streaming logic.

Use it for:

- Expose LangGraph agent workflows as HTTP endpoints without manual endpoint scaffolding.
- Stream agent thinking, tool calls, and state updates to a frontend using the AG-UI event protocol.
- Convert between LangChain message formats and AG-UI protocol automatically during request/response cycles.
- Build agent UI applications that display real-time agent execution events and state changes.

## Worth the install?

AI-flagged interpretation of the facts above — verify before relying.

Bridges LangGraph agent workflows to the AG-UI protocol, automatically generating HTTP endpoints with event streaming for agent interactions.

Yes, if you are building a web-based agent UI on top of LangGraph and want to adopt the AG-UI protocol standard. The package eliminates boilerplate for endpoint creation and event streaming. Install with confidence: MIT license, active maintenance, no known vulnerabilities, and low dependency friction. Not necessary if you are not using LangGraph or the AG-UI protocol.

## Install

pip install ag-ui-langgraph
uv add ag-ui-langgraph
poetry add ag-ui-langgraph

## Installing ag-ui-langgraph

Before you install: Low install friction with a pure-wheel distribution. Marked active with a recent release cycle (56 days since last update). Depends on established packages: langchain, langgraph, pydantic, and ag-ui protocol libraries.

License in practice: MIT license permits commercial and private use with minimal restrictions, making it suitable for most projects.

Quickstart:

from ag_ui_langgraph import add_langgraph_fastapi_endpoint

add_langgraph_fastapi_endpoint(app, graph, "/agent")

Requires Python 3.10 or later (supports up to <3.15); a LangGraph workflow graph object must be provided.

Verify before relying:

- Whether the package handles concurrent agent requests or has scaling limitations
- What AG-UI protocol versions are supported and whether breaking changes are documented
- Whether message translation between AG-UI and LangChain formats covers all message types
- Whether the endpoint creation requires a specific web framework or is framework-agnostic

## Package facts

- License: MIT (permissive)
- Python support: supports_current
- Install friction: low
- Maintenance: active
- Downloads: 801.3K/month (top 15,000 on PyPI)
- Known vulnerabilities: none known

## Tags

langgraph http endpoint, ag-ui protocol implementation, langchain agent streaming, langgraph event streaming, agent ui protocol bridge, langgraph-integration, agent-framework, protocol-bridge

[View on SkillFed](https://skillfed.io/packages/ag-ui-langgraph) · [View on PyPI](https://pypi.org/project/ag-ui-langgraph/)