NouveauVoie Eco désormais sur 91 modèles : le même modèle, moins cher

Devise

Vous rechargez toujours en euros. Les autres devises sont affichées au taux du jour de la BCE. Taux de référence de la BCE

Se connecterCommencer

Documentationv1

Menu de la documentationLangChain

LangChain

Use Nexrouting models in LangChain for Python and JavaScript through the OpenAI integration.

La documentation est en anglais.

LangChain's OpenAI integration works with Nexrouting. Set the base URL to https://api.nexrouting.com/v1, use your Nexrouting key, and pass any chat model ID, from any provider.

Python

Shell
pip install langchain-openai
Python
import osfrom langchain_openai import ChatOpenAI llm = ChatOpenAI(    model="anthropic/claude-sonnet-5.5",    base_url="https://api.nexrouting.com/v1",    api_key=os.environ["NEXROUTING_API_KEY"],)print(llm.invoke("Hello").content)

JavaScript

Shell
npm install @langchain/openai
Node
import { ChatOpenAI } from "@langchain/openai"; const llm = new ChatOpenAI({  model: "openai/gpt-6-astra",  apiKey: process.env.NEXROUTING_API_KEY,  configuration: { baseURL: "https://api.nexrouting.com/v1" },});const reply = await llm.invoke("Hello");console.log(reply.content);

Embeddings

Turn off check_embedding_ctx_length, so LangChain sends plain text instead of token IDs counted with an OpenAI tokenizer.

Python
import osfrom langchain_openai import OpenAIEmbeddings embeddings = OpenAIEmbeddings(    model="openai/text-embedding-3-small",    base_url="https://api.nexrouting.com/v1",    api_key=os.environ["NEXROUTING_API_KEY"],    check_embedding_ctx_length=False,)vector = embeddings.embed_query("Hello")

Streaming, tool calling and structured output work as with OpenAI, for models that support them.