DEV Community

APIVAI
APIVAI

Posted on • Originally published at apivai.com

Using an OpenAI-Compatible API with n8n, LangChain & LlamaIndex

#n8n #langchain #integration

One base URL change, three frameworks

n8n, LangChain, and LlamaIndex all speak the OpenAI-compatible API. That means you can point them at any OpenAI-compatible gateway by changing the base URL and API key — no custom integration code. This guide shows the exact setting in each, using APIVAI (which resells Claude and GPT at a steep discount) as the endpoint.

The APIVAI base URL is https://api.apivai.com/v1. Pick a model from /v1/models before hardcoding a name.

n8n

In n8n, the OpenAI nodes (and the AI Agent / Chat Model nodes) accept a custom credential. Create an "OpenAI" credential and set:

  • API Key: your APIVAI key
  • Base URL: https://api.apivai.com/v1

Then choose a model ID returned by /v1/models in the node. Existing n8n OpenAI nodes and agent workflows keep working — only the credential changed. This is the fastest way to run cheaper models inside automation flows.

LangChain (Python)

LangChain's ChatOpenAI takes a base_url and api_key. Set them and you are done.

from langchain_openai import ChatOpenAI
llm = ChatOpenAI(base_url="https://api.apivai.com/v1", api_key="YOUR_APIVAI_API_KEY", model="claude-sonnet-4-6")
print(llm.invoke("Write one sentence about API gateways.").content)
Enter fullscreen mode Exit fullscreen mode

The same pattern works for chains, agents, and tool calling — anything built on ChatOpenAI inherits the endpoint.

LangChain (JavaScript) 

import { ChatOpenAI } from "@langchain/openai";
const llm = new ChatOpenAI({ model: "claude-sonnet-4-6", apiKey: "YOUR_APIVAI_API_KEY", configuration: { baseURL: "https://api.apivai.com/v1" } });
const res = await llm.invoke("Write one sentence about API gateways.");
console.log(res.content);
Enter fullscreen mode Exit fullscreen mode

LlamaIndex (Python)

LlamaIndex's OpenAI-compatible LLM wrapper accepts an api_base.

from llama_index.llms.openai_like import OpenAILike
llm = OpenAILike(model="claude-sonnet-4-6", api_base="https://api.apivai.com/v1", api_key="YOUR_APIVAI_API_KEY", is_chat_model=True)
print(llm.complete("Write one sentence about API gateways."))
Enter fullscreen mode Exit fullscreen mode

Verify before you build

Whatever framework you use, run a tiny request first to confirm the endpoint, key, and model are wired correctly. A 401 means the key is not being read; a 404 usually means the base URL is missing /v1; a model error means the configured name is not in the current /v1/models response.

Get started

Create an APIVAI key, drop the base URL into your n8n credential or your ChatOpenAI / OpenAILike config, pick a model from /v1/models, and your existing framework code runs against cheaper models with no rewrite.

Top comments (0)