MuiRouter

OpenAI / Claude / DeepSeek / GLM / Qwen / Grok / Kimi / MiniMax — one router

One AI API Router for Every LLM

MuiRouter is an OpenAI-compatible API router: one key routes your requests to OpenAI, Claude, DeepSeek, GLM, Qwen, Kimi, Grok, MiniMax and more. No VPN, no overseas account — top up and pay at list price.Drop-in compatible with the OpenAI SDK — just change the Base URL.

9
AI Providers
100%
OpenAI Compatible
300+
Cloudflare Nodes
$0
Monthly / Subscription Fee

Route every major LLM through one API

Supporting OpenAI, Anthropic, DeepSeek, Moonshot AI, xAI, Zhipu GLM, Qwen, MiniMax, Meta and more — continuously expanding

OpenAI

The most powerful general model — code, reasoning, multimodal

GPT-6 AstraGPT-5.6 SolGPT-5.6 TerraGPT-5.6 LunaGPT Image 2
Anthropic

Ultra-long context, excellent coding and analysis

Claude Fable 5.1Claude Opus 5Claude Sonnet 5Claude Haiku 4.5Claude Opus 4.6
DeepSeek

DeepSeek V4 Flash with 284B MoE architecture, 1M-token context, and ultra-low prompt caching costs

DeepSeek V4.1 FlashDeepSeek V4 FlashDeepSeek V4 ProDeepSeek V3DeepSeek R1
Moonshot AI

Kimi K3 with a 1M-token context, vision, always-on thinking, and prompt caching

Kimi K3Kimi K2.7 CodeKimi K2.6Kimi K2
xAI

Opus-class reasoning and coding, with a 1M-token context option

Grok 4.6Grok 4.5Grok 4.3Grok Imagine Image
Zhipu GLM

Zhipu GLM flagship reasoning and coding models at list price

GLM-5.3GLM-5.3 FlashGLM-5.2GLM-4.7 Flash
Qwen

Alibaba Qwen Max/Plus and Qwen3 MoE models at list price

Qwen3.8 MaxQwen3.7 MaxQwen3.7 PlusQwen3-30B
MiniMax

MiniMax M2/M3 series for fast and cost-effective inference

MiniMax M3MiniMax M2.7MiniMax M2.5
Meta

Meta in-house Muse Spark model, lightweight and cost-effective

Muse Spark 1.3Muse Spark 1.2
View all models →

Supports both OpenAI-compatible mode and native API passthrough — no features lost

Image API

Text-to-image and image editing are on the same gateway

Call GPT Image models through the same API key and billing pipeline. Generate new assets from prompts, or upload reference images for edits without a separate integration.

POST /v1/images/generations

POST /v1/images/edits

Generate from prompt

Use /v1/images/generations for product shots, diagrams, social assets, and visual drafts.

Edit uploaded images

Use /v1/images/edits with multipart uploads to transform, combine, or refine existing images.

One account, one bill

Image requests share the same model registry, wallet balance, usage logs, and API key flow as chat models.

Why developers choose MuiRouter

The AI API router and LLM gateway built for developers worldwide

No VPN Required

Proxied through Cloudflare's global edge network for barrier-free access. No overseas credit card or foreign account needed.

Pay Per Use

No monthly fees, no subscriptions. Top up your wallet, pay only for what you use. Balance never expires, check spending anytime.

OpenAI Compatible

Fully compatible with OpenAI SDK and API format. Existing projects only need to change Base URL and API Key — zero migration cost.

Native API Passthrough

Also supports native APIs from each provider, preserving all parameters and features. Claude's long context, Gemini's multimodal, Grok's agentic coding — nothing missed.

Get Started in 3 Steps

From sign-up to API call in as fast as 1 minute

STEP 01

Create Account

Email registration, done in 30 seconds. Get your personal API Key instantly.

STEP 02

Try First, Then Top Up

Start with a small top-up, then scale as you need.

STEP 03

Start Calling

Replace the Base URL — no other code changes needed.

Integration Is This Simple

Already using OpenAI SDK? Just change two lines

example.py
from openai import OpenAI

client = OpenAI(
    api_key="sk-your-mui-router-key",
    base_url="https://your-domain.com/v1",
)

# It's that simple — exactly like using OpenAI directly
response = client.chat.completions.create(
    model="gpt-4o",          # or claude-sonnet, gemini-pro
    messages=[{"role": "user", "content": "Hello"}],
)

print(response.choices[0].message.content)

Also supports Node.js, curl, and any OpenAI API-compatible tool or framework

Or run Claude Code on your balance

Point Claude Code or the Anthropic SDK at MuiRouter — no Anthropic account needed

terminal
# Point Claude Code at MuiRouter — native /v1/messages
export ANTHROPIC_BASE_URL=https://your-domain.com
export ANTHROPIC_AUTH_TOKEN=sk-your-mui-router-key

claude   # runs Claude on your MuiRouter balance

Works with the Anthropic SDK too — set ANTHROPIC_BASE_URL and use your sk-gw- key.

Frequently asked questions

Is MuiRouter related to Material UI (MUI)?
No. MuiRouter is an OpenAI-compatible AI/LLM API router and MCP server, named after the muirouter.com domain. It has nothing to do with the Material UI (MUI) React component library or MUI X.
What is MuiRouter?
MuiRouter is an AI API router: one key routes your requests to OpenAI, Anthropic Claude, Google Gemini, xAI Grok, Cloudflare Workers AI and Xiaomi MiMo, billed per token with no subscription. Built by developers for ourselves and friends, it provides a dependable endpoint and convenient infrastructure for future AI applications.
Is it OpenAI-compatible?
Yes. Point the OpenAI SDK's Base URL at MuiRouter and keep your existing code. See the OpenAI-compatible router page to learn more.
Can I use Claude Code or the Anthropic SDK with MuiRouter?
Yes. MuiRouter exposes a native Anthropic /v1/messages endpoint. Set ANTHROPIC_BASE_URL to MuiRouter and use your sk-gw- key — Claude Code, the Anthropic SDK and any Anthropic-native client run on your MuiRouter balance, with no Anthropic account required.