Quickstart

Code that calls OpenRouter calls InferFlux after two changes: the base URL and the key. The model slugs, the request and response shapes, streaming, errors and the provider routing object stay as they are.

The two changes

Base URL: https://testnet.inferflux.xyz/api/v1
Key: One of your keys, from the keys page, kept in INFERFLUX_API_KEY. Keys begin with sk-ifx-, so secret scanners still find them.

Before and after

Lines marked with a minus are OpenRouter's; lines marked with a plus replace them. The rest does not change.

Python, the OpenAI SDK

 import os
 from openai import OpenAI
 
 client = OpenAI(
-    base_url="https://openrouter.ai/api/v1",
-    api_key=os.environ["OPENROUTER_API_KEY"],
+    base_url="https://testnet.inferflux.xyz/api/v1",
+    api_key=os.environ["INFERFLUX_API_KEY"],
 )
 
 completion = client.chat.completions.create(
     model="deepseek/deepseek-v4.1-flash-1017",
     messages=[{"role": "user", "content": "Hello"}],
 )

TypeScript, the OpenAI SDK

 import OpenAI from "openai";
 
 const client = new OpenAI({
-  baseURL: "https://openrouter.ai/api/v1",
-  apiKey: process.env.OPENROUTER_API_KEY,
+  baseURL: "https://testnet.inferflux.xyz/api/v1",
+  apiKey: process.env.INFERFLUX_API_KEY,
 });
 
 const completion = await client.chat.completions.create({
   model: "deepseek/deepseek-v4.1-flash-1017",
   messages: [{ role: "user", content: "Hello" }],
 });

curl

-curl https://openrouter.ai/api/v1/chat/completions \
-  -H "Authorization: Bearer $OPENROUTER_API_KEY" \
+curl https://testnet.inferflux.xyz/api/v1/chat/completions \
+  -H "Authorization: Bearer $INFERFLUX_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{"model": "deepseek/deepseek-v4.1-flash-1017", "messages": [{"role": "user", "content": "Hello"}]}'

What stays the same

Model slugs, as the model list shows them, with :free for free endpoints.
The provider object: order, only, ignore, allow_fallbacks, require_parameters, data_collection, zdr, quantizations and sort, plus two of ours: class and min_score.
Streaming, with usage at the end when the request includes it, and the error shapes.
The HTTP-Referer and X-Title headers that name your app.

What you get besides

Each provider's measured latency, throughput, uptime and floor score on every model's page, beside its price.
A receipt for every request, in Logs, and a report button that refunds a generation the validators find was not the model.
A balance in USDC that is yours, with limits on each key that the chain enforces.

In a framework

The same two changes, in three common ones.

LangChain, Python

import os
from langchain_openai import ChatOpenAI

llm = ChatOpenAI(model="deepseek/deepseek-v4.1-flash-1017", base_url="https://testnet.inferflux.xyz/api/v1", api_key=os.environ["INFERFLUX_API_KEY"])

LiteLLM

import os
import litellm

response = litellm.completion(
    model="openai/deepseek/deepseek-v4.1-flash-1017",
    api_base="https://testnet.inferflux.xyz/api/v1",
    api_key=os.environ["INFERFLUX_API_KEY"],
    messages=[{"role": "user", "content": "Hello"}],
)

Vercel AI SDK

import { createOpenAI } from "@ai-sdk/openai";
import { generateText } from "ai";

const inferflux = createOpenAI({ baseURL: "https://testnet.inferflux.xyz/api/v1", apiKey: process.env.INFERFLUX_API_KEY });
const { text } = await generateText({ model: inferflux.chat("deepseek/deepseek-v4.1-flash-1017"), prompt: "Hello" });