Build LangChain agents with Niuron.

Use ChatOpenAI with Niuron’s API base and agent routing preset.

Chat Completions

Set up the connection Create an account ↗

YOUR TOOLSAPI FORMATS

Three HTTP API formats, connected through one Niuron key.

01 / CONNECTION GUIDE

Connect in three steps

Create your key
Create a Niuron key and export NIURON_API_KEY.
Connect your tool
Install langchain-openai in your Python environment.
Choose your model
Run the example with niuron/agent and use_responses_api=False; add tools after a successful text request.

02 / Before your first request

Try a small request first

Send a short prompt, check the response, then confirm the request and cost in your Niuron dashboard. Agent loops make multiple requests; set an API key budget before running unattended tasks.

What this setup supports

This example uses Chat Completions. Niuron’s /responses and /messages endpoints are also available to suitable clients; they do not enable every LangChain hosted-tool, state or embedding feature.

Official configuration reference ↗

niuron.pyLangChain

Connection settings

import os
from langchain_openai import ChatOpenAI

model = ChatOpenAI(
    model="niuron/agent",
    api_key=os.environ["NIURON_API_KEY"],
    base_url="https://api.niuron.ai/v1",
    use_responses_api=False,
)
print(model.invoke("Hello, Niuron!").content)

Replace placeholders or set NIURON_API_KEY on your server or shell before running the example.

Your next request starts here.

Create a Niuron key, connect your tool, and see every request in your dashboard.

Create an account ↗Estimate your token costs →

Build LangChain agents with Niuron.

Connect in three steps

Create your key

Connect your tool

Choose your model

Try a small request first

Your next request starts here.