LangGraph OpenAI Serve

Serve LangGraph graphs through an OpenAI-compatible /v1 API so existing OpenAI SDKs, Chainlit, Open WebUI, and similar clients can call them without a project-specific protocol.

Install

Use uv for project dependency management:

uv add langgraph-openai-serve

The equivalent pip command is:

pip install langgraph-openai-serve

The package contains the OpenAI-compatible server integration, not a built-in LLM graph. Applications register their own graphs. The demo/ checkout keeps each deployable application in an independent uv project with its own lockfile. Its lgos-rag example indexes a small corpus packaged with the demo API, so the entire directory can be copied and run without files from this repository.

Quick Demo

From this repository, prepare the demo environment and PostgreSQL:

cd demo
cp .env.example .env
docker compose -f compose.yaml up -d lgos-db
uv run --directory api --env-file ../.env \
  --locked --with-editable ../.. lgos-demo-api-setup
uv run --directory api --env-file ../.env \
  --locked --with-editable ../.. lgos-demo-api

Then call the demo with the OpenAI Python client:

from openai import OpenAI

client = OpenAI(base_url="http://localhost:8000/v1", api_key="DUMMY")

response = client.chat.completions.create(
    model="custom-input-output-context",
    messages=[{"role": "user", "content": "Show me custom schemas."}],
    user="demo-user",
)

print(response.choices[0].message.content)

Use curl http://localhost:8000/v1/models only as a diagnostic to inspect the registered demo graph names.

The optional editable overlay tests this checkout without changing the self-contained demo project or its lockfile. The demo publishes independent API and Chainlit images and uses official images for third-party services such as Open WebUI. See the demo Docker Compose guide.

The demo also includes a PostgreSQL-persistent Chainlit client. It uses a shared mock login by default, with PocketID OAuth available as an opt-in mode. See the Chainlit demo.

Use In FastAPI

from fastapi import FastAPI
from langgraph_openai_serve import GraphConfig, GraphRegistry, LanggraphOpenaiServe
from your_graphs import my_graph

app = FastAPI()
graphs = GraphRegistry(
    registry={
        "my-graph": GraphConfig(
            graph=my_graph,
            description="Answer questions with my LangGraph workflow.",
            streamable_node_names=["generate"],
        )
    }
)

LanggraphOpenaiServe(app=app, graphs=graphs).bind_openai_api()

The default base URL is {host}/v1. Registered graph names become OpenAI model values.

Docs

• Documentation home: docs/index.md
• Package getting started: docs/getting-started.md
• Self-contained demo stack: docs/demo/index.md
• Runnable demo API: docs/demo/api.md
• Demo graph catalog: docs/demo/graphs.md
• OpenAI clients: docs/tutorials/openai-clients.md
• Custom graphs: docs/tutorials/custom-graphs.md
• LangGraph runtime settings: docs/how-to-guides/langgraph-runtime-settings.md
• OpenAI-compatible proxies: docs/how-to-guides/openai-proxies.md
• API and configuration: docs/reference.md
• Compatibility contract: docs/explanation/openai-compatibility.md