---
title: LangChain
description: Connect Clipia AI Gateway to LangChain via ChatOpenAI with the Clipia base_url — chat, tools and tool agents on Claude, GPT and Gemini.
---

Clipia AI Gateway is OpenAI-contract compatible, so it plugs into LangChain through the standard `ChatOpenAI` class — just point it at our `base_url` with a Clipia key. From there, all the usual LangChain features work: chat, tools, agents and structured output.

## Initialization

<Tabs items={['Python', 'Node.js']}>
<Tab value="Python">

```bash
pip install langchain-openai
```

```python
import os
from langchain_openai import ChatOpenAI

llm = ChatOpenAI(
    model="claude-opus-5",
    base_url="https://api.clipia.ai/v1",
    api_key=os.environ["CLIPIA_API_KEY"],
    temperature=0.7,
)

print(llm.invoke("Tell me three facts about Mars.").content)
```

</Tab>
<Tab value="Node.js">

```bash
npm install @langchain/openai @langchain/core
```

```ts
import { ChatOpenAI } from "@langchain/openai";

const llm = new ChatOpenAI({
  model: "claude-opus-5",
  apiKey: process.env.CLIPIA_API_KEY,
  configuration: { baseURL: "https://api.clipia.ai/v1" },
  temperature: 0.7,
});

const res = await llm.invoke("Tell me three facts about Mars.");
console.log(res.content);
```

</Tab>
</Tabs>

<Callout type="info" title="One parameter to switch provider">
If you already have `ChatOpenAI` code, switching over just means setting `base_url` (`configuration.baseURL` in Node.js) and a Clipia key. The `model` field is a string from the [catalog](/en/docs/llm-gateway/models).
</Callout>

## Tool agent

Bind tools via `bind_tools`, or assemble a ready-made ReAct agent with `langgraph`. Example with a single function tool:

```python
import os
from langchain_openai import ChatOpenAI
from langchain_core.tools import tool
from langgraph.prebuilt import create_react_agent

@tool
def get_weather(city: str) -> str:
    """Return the current weather in a city."""
    return f"In {city}: 14°C, cloudy"

llm = ChatOpenAI(
    model="claude-opus-5",
    base_url="https://api.clipia.ai/v1",
    api_key=os.environ["CLIPIA_API_KEY"],
)

agent = create_react_agent(llm, tools=[get_weather])

result = agent.invoke(
    {"messages": [{"role": "user", "content": "What's the weather in Moscow?"}]}
)
print(result["messages"][-1].content)
```

The agent runs the full loop on its own: requests a tool call, gets the result and produces the final answer.

## Structured output

`with_structured_output` uses `response_format` under the hood — the model returns data strictly matching the schema (a Pydantic model or a JSON schema):

```python
from pydantic import BaseModel

class Person(BaseModel):
    name: str
    age: int

structured = llm.with_structured_output(Person)
person = structured.invoke("Extract name and age: Anna is 30.")
print(person.name, person.age)  # Anna 30
```

<Callout type="info" title="Streaming">
Streaming is available via `llm.stream(...)` (Python) and `await llm.stream(...)` (Node.js) — the gateway emits SSE chunks in OpenAI format. More on streaming and tools on the [Chat Completions](/en/docs/llm-gateway/chat-completions) page.
</Callout>