API documentation

Build with the InferGate API

Use an OpenAI-compatible base URL and keep every API Key on your server. Start by asking the API which models your account can access.

Base URL

https://api.useinfergate.com/v1

Keep keys server-side.

Never embed a complete InferGate API Key in browser code, mobile bundles, screenshots, support messages, or source control.

1. Get the account model list

Call this endpoint first and select only a returned model.

cURL
curl https://api.useinfergate.com/v1/models \
  -H "Authorization: Bearer $INFERGATE_API_KEY"

2. Responses API

cURL
curl https://api.useinfergate.com/v1/responses \
  -H "Authorization: Bearer $INFERGATE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model":"gpt-5.4-mini","input":"Reply with: ready"}'
JavaScript
import OpenAI from "openai";

const client = new OpenAI({
  apiKey: process.env.INFERGATE_API_KEY,
  baseURL: "https://api.useinfergate.com/v1",
});

const response = await client.responses.create({
  model: "gpt-5.4-mini",
  input: "Reply with: ready",
});

console.log(response.output_text);
Python
import os
from openai import OpenAI

client = OpenAI(
    api_key=os.environ["INFERGATE_API_KEY"],
    base_url="https://api.useinfergate.com/v1",
)

response = client.responses.create(
    model="gpt-5.4-mini",
    input="Reply with: ready",
)

print(response.output_text)

3. Chat Completions

cURL
curl https://api.useinfergate.com/v1/chat/completions \
  -H "Authorization: Bearer $INFERGATE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model":"gpt-5.4-mini","messages":[{"role":"user","content":"Reply with: ready"}]}'

Set "stream": true to receive a server-sent event stream.

4. Image generation

gpt-image-2 requires paid model access.

cURL
curl https://api.useinfergate.com/v1/images/generations \
  -H "Authorization: Bearer $INFERGATE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model":"gpt-image-2","prompt":"A precise black ink diagram of a network gateway on white paper"}'

Access and availability errors

HTTPCodeMeaning
403access_deniedThe account has not completed a qualifying real payment for this model.
503model_unavailableThe requested model is explicitly unavailable; do not silently substitute it.

Paid model access remains after the paid balance reaches zero. A full refund or chargeback can revoke it when no other valid payment remains. InferGate is independent from OpenAI and other model providers.