Xyrus RouterAI gateway

One key. Every model.

A single OpenAI-compatible gateway to OpenAI, and more. Build once, route anywhere.

Get your API key Read the docs
18 modelsNative streamingNo vendor lock-in
chat.ts
const response = await xyrus.chat.completions.create({model: "claude-sonnet-4-6", messages: [{role: "user", content: "Ship faster."}], stream: true});
Routing request to Anthropic · 142ms
OpenAI✦supported provider
Anthropic✦supported provider
Google✦supported provider
DeepSeek✦supported provider
xAI✦supported provider
Qwen✦supported provider
OpenAI✦supported provider
Anthropic✦supported provider
Google✦supported provider
DeepSeek✦supported provider
xAI✦supported provider
Qwen✦supported provider

Why Xyrus

The boring parts of AI infrastructure, handled.

Multi-model

Every provider, one interface

Switch between 13+ models across 6 providers without changing integration code.

Drop-in

OpenAI-compatible by default

Change the base URL and your existing SDKs keep working.

Transparent

Know what every request costs

Token usage and cost attribution, visible per model and API key.

Reliable

Limits that stay predictable

Sliding-window rate limits protect your quotas and budgets.

Real-time

Stream every response

Native SSE streaming across all supported providers.

Lean runtime

Deploy without the baggage

Node.js and SQLite. No Redis, Postgres, or managed service required.

Model registry

Choose the model, not the integration.

One consistent request shape. Eight providers. No context switching.

GPT-5.6

xyrz/gpt-5.6-luna · xyrz/gpt-5.6-sol · xyrz/gpt-5.6-terra

03 models

GPT-5.5

xyrz/gpt-5.5 · xyrz/gpt-5.5-review

02 models

MiniMax

xyrz/minimax-m2.1 · xyrz/minimax-m3

02 models

GLM

xyrz/glm-5.1

01 model

Kimi

xyrz/kimi-k2.7-code · xyrz/kimi-k3

02 models

DeepSeek

xyrz/deepseek-v4-flash · xyrz/deepseek-v4-pro

02 models

Claude

xyrz/claude-fable-5 · xyrz/claude-opus-4.7 · xyrz/claude-opus-5

03 models

Qwen

xyrz/qwen3.7-max · xyrz/qwen3.7-plus · xyrz/qwen3.8-max-preview

03 models

13+

models available

6

AI providers

1

unified endpoint

99.9%

gateway uptime target

Pricing

Top up token, tanpa ribet.

Kuota dihitung sebagai (input + output) × multiple model. Paket kuota tidak pernah kadaluarsa; paket unlimited berlaku sesuai durasinya.

Top-up kuota

3.5 Miliar Token

Rp150.000

Paket starter untuk eksperimen dan aplikasi kecil.

Beli paket

10 Miliar Token

Rp200.000

Cocok untuk aplikasi produksi skala menengah.

Beli paket

15 Miliar Token

Rp250.000

Nilai terbaik untuk tim kecil yang aktif.

Beli paket

25 Miliar Token

Rp350.000

Untuk aplikasi dengan traffic harian tinggi.

Beli paket

40 Miliar Token

Rp500.000

Untuk produk yang melayani banyak pengguna.

Beli paket

250 Miliar Token

Rp1.500.000

Kuota besar tanpa kadaluarsa untuk skala enterprise.

Beli paket
Paket unlimited

Unlimited 1 Minggu

Unlimited

Rp550.000

Pemakaian tanpa batas selama 7 hari.

Aktifkan

Unlimited 1 Bulan

Unlimited

Rp2.500.000

Pemakaian tanpa batas selama 30 hari.

Aktifkan

FAQ

Before you ship.

An API gateway that lets you access multiple AI models through a single OpenAI-compatible endpoint and API key.

One integration, one API key, one set of error handling. Switch models without changing code and track usage across all providers in one dashboard.

13+ models including GPT-4o, GPT-4.1, Claude Sonnet 4, Claude Opus 5, Gemini 2.5 Pro, DeepSeek V4, Grok 4.5, and Qwen 3.7.

Yes. SSE streaming is supported across all providers. Add stream: true just like OpenAI's API.

Yes. Xyrus Router runs as a Node.js process with SQLite. No external infrastructure is required.

Start routing

The next model is one API call away.

Get your key, point your SDK at Xyrus Router, and keep shipping.

Create your key