Apifuse — OpenAI-Compatible API Gateway (GPT-5, Claude, Gemini)

OpenAI-compatible API for GPT-5, Claude and Gemini. Multi-provider routing, transparent pricing, low latency. Enterprise SLA, SDK support, pay-as-you-go.

One unified OpenAI-compatible endpoint for GPT-5, Claude and Gemini. Migrate from OpenAI in minutes by simply changing your base URL to https://api.apifuse.net/v1 — keep your existing SDKs with no code rewrite. Multi-provider routing ensures low latency and high uptime, with transparent pricing, an enterprise SLA and global CDN acceleration.

Quick Start

Create an API key, point your client at the Apifuse base URL, and call any of the 500+ models from the Model Market.

Text Series

Every text endpoint mirrors a provider's native request and response shape, so existing tooling, SDKs and prompts work unchanged. Pick the schema you already use — OpenAI, Anthropic or Google — and only the base URL changes.

General Chat API (Streaming)

Standard /v1/chat/completions with "stream": true. Tokens arrive as server-sent events for low time-to-first-token. Works with GPT-5, Claude and Gemini through the same OpenAI schema.

# Streaming chat completion — OpenAI-compatible
curl https://api.apifuse.net/v1/chat/completions \
  -H "Authorization: Bearer <token>" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-5",
    "messages": [{ "role": "user", "content": "Hello" }],
    "stream": true
  }'

General Chat API (Non-Streaming)

The same /v1/chat/completions endpoint with "stream": false returns one complete JSON response — the simplest option for batch jobs and server-to-server calls.

# Non-streaming chat completion
curl https://api.apifuse.net/v1/chat/completions \
  -H "Authorization: Bearer <token>" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-5",
    "messages": [{ "role": "user", "content": "Hello" }],
    "stream": false
  }'

Claude Messages API

Call Claude models with Anthropic's native /v1/messages schema — system prompts, multi-turn messages and max_tokens behave exactly as they do against Anthropic directly.

# Claude Messages API — Anthropic-native schema
curl https://api.apifuse.net/v1/messages \
  -H "Authorization: Bearer <token>" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "claude-sonnet-4-5",
    "max_tokens": 1024,
    "messages": [{ "role": "user", "content": "Hello" }]
  }'

Gemini Native Format

Use Google's native /v1beta/models/{model}:generateContent format with contents and parts — ideal if your code already targets the Gemini SDK.

# Gemini native generateContent
curl https://api.apifuse.net/v1beta/models/gemini-2.5-pro:generateContent \
  -H "Authorization: Bearer <token>" \
  -H "Content-Type: application/json" \
  -d '{
    "contents": [{ "parts": [{ "text": "Hello" }] }]
  }'

OpenAI Multimodal Responses API

The newer /v1/responses endpoint accepts mixed text and image input and returns structured multimodal output — a drop-in for OpenAI's Responses API.

# OpenAI Responses API — multimodal input
curl https://api.apifuse.net/v1/responses \
  -H "Authorization: Bearer <token>" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-5",
    "input": "Describe this product in one line."
  }'

Image Series

Generate and edit images via /v1/images/generations across Nano Banana, GPT Image and Imagen — one schema, per-resolution pricing tiers, and billing only on successful renders. Switch model by changing the model field.

Nano Banana 2

Google's gemini-3-pro-image-preview (alias nano-banana-2) for high-fidelity 1K–4K generation with strong text rendering.

# Image generation — Nano Banana 2
curl https://api.apifuse.net/v1/images/generations \
  -H "Authorization: Bearer <token>" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "nano-banana-2",
    "prompt": "A red panda barista, studio lighting",
    "size": "1024x1024"
  }'

GPT Image 2

OpenAI's gpt-image-2 — near-perfect text rendering and native 4K output. Pass "model": "gpt-image-2" to the same generations endpoint.

Imagen 4.0

Google DeepMind's flagship text-to-image model — up to 2K resolution with improved prompt adherence. Use "model": "imagen-4.0".

Video Series

Asynchronous video generation via /v1/videos for Sora, Veo and Kling. Submit a job, poll the returned id for completion, and pay only for successful renders.

Sora 2 (Text-to-Video)

OpenAI's sora-2 for cinematic motion, synced audio and up to 20-second 1080p clips. Submit a render job, then poll for the result URL.

# Submit a Sora 2 render job
curl https://api.apifuse.net/v1/videos \
  -H "Authorization: Bearer <token>" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "sora-2",
    "prompt": "A timelapse of a city at dusk",
    "duration": 5,
    "resolution": "1080p"
  }'

# Poll for completion
curl https://api.apifuse.net/v1/videos/<id> \
  -H "Authorization: Bearer <token>"

Veo 3.1

Google DeepMind's veo-3.1 — low-latency generation with strong physics and prompt fidelity. Same submit-and-poll flow with "model": "veo-3.1".

Kling V3

Kuaishou's kling-v3 — high-motion video with extended duration and precise camera control. Use "model": "kling-v3" on the same endpoint.

Platform Benefits

SDKs & Code Examples

Use the official OpenAI SDKs — only the base URL changes.

Python SDK

from openai import OpenAI

client = OpenAI(
    api_key="<token>",
    base_url="https://api.apifuse.net/v1",
)

resp = client.chat.completions.create(
    model="gpt-5",
    messages=[{"role": "user", "content": "Hello"}],
)
print(resp.choices[0].message.content)

Node.js SDK

import OpenAI from "openai";

const client = new OpenAI({
  apiKey: "<token>",
  baseURL: "https://api.apifuse.net/v1",
});

const resp = await client.chat.completions.create({
  model: "gpt-5",
  messages: [{ role: "user", content: "Hello" }],
});
console.log(resp.choices[0].message.content);