THE UNIVERSAL AI GATEWAY

One endpoint for
every frontier model.

Low-latency, ban-safe access to Claude, GPT and Gemini — through a single drop-in API. Usage-based billing, balance that never expires, Alipay & WeChat Pay supported.

99.98%
Uptime SLA
38ms
Median edge latency
4
Frontier providers
omnirouter — quickstart
$ curl https://omnirouter.cc/v1/messages \
  -H "authorization: Bearer $OMNI_KEY" \
  -d '{
    "model": "claude-opus-4-8",
    "max_tokens": 1024,
    "messages": [{ "role": "user", "content": "Ship it." }]
  }'
routed via edge-sg-3
200 OK · 38ms

Unified access to

Claude Opus 4.8GPT-5.2 CodexGemini 3.0 ProClaude Sonnet 4.6o5-miniGemini Flash
Ecosystem

Plug into the tools you already ship with

Drop OmniRouter behind your favourite AI coding agents — one key powers them all.

OpenClaw

A local AI agent that runs on your own machine — executing real tasks through chat, not just conversation.

OPEN SOURCE · LOCAL

Claude Code

Anthropic's official CLI with native Extended Thinking — coding that feels like a conversation.

BY ANTHROPIC

Codex

OpenAI's coding agent — large-scale refactors, bug fixing, and test generation, stable on long tasks.

BY OPENAI

Gemini CLI

Google's open-source terminal agent — invoke Gemini for coding, debugging, and workflow automation.

BY GOOGLE

Pricing

Native models, exclusive rates

1 RMB = 1 USD, billed to the token. No subscription, no lock-in, no expiry.

PAYGO

Pay As You Go

1 ¥ = 1 $

No subscription. Top up and spend across every model.

  • Equivalent balance on top-up
  • Pay only for what you use
  • Balance never expires
Top up now
Most popular

RECOMMENDED

Claude On-Demand

from $1.79 / M

Usage-based access tuned for Claude Code workflows.

  • Synced with official pricing
  • All Claude models
  • Optimized for Claude Code
Get started

RECOMMENDED

Codex On-Demand

from $0.75 / M

Flexible, usage-based billing for OpenAI agents.

  • Synced with official pricing
  • All GPT & Codex models
  • Optimized for Codex
Get started
Why OmniRouter

Infrastructure that gets out of your way

Let top-tier AI write the code — we handle the routing, the failover, and the bill.

Global Anycast Edge

Stable access from anywhere, with intelligent routing that trims latency and survives network jitter.

Automatic Failover

Distributed upstream pools with health-aware switching — reliable exactly when it matters.

One-Line Integration

Swap the base URL. No rewrites — fully OpenAI & Anthropic wire-compatible out of the box.

Zero Ban Risk

Rate-isolated brokering means your traffic never touches a personal account. Run agents freely.

Token-Level Billing

Pay only for what you use, down to the request. Balance never expires — top up once, spend anytime.

Every Frontier Model

Claude, GPT, Gemini — one unified key, synced to official capabilities within hours of release.

99.98%
Uptime SLA
38ms
Median edge latency
4
Frontier providers
12B+
Tokens routed / day
FAQ

Frequently asked questions

No. OmniRouter brokers requests through dedicated, rate-isolated upstream pools. Your traffic never touches a personal account, so there is zero ban risk — even under heavy agentic workloads.

Every frontier model worth shipping on: Claude (Opus / Sonnet / Haiku), OpenAI GPT & Codex, and Google Gemini. New flagships are added within hours of release.

Volume aggregation plus intelligent caching and routing. Group multipliers are synced to official pricing, then discounted — you pay a fraction of list price with no monthly commitment.

Never. Top up once and the balance stays until you spend it. Pay only for the tokens you actually use, down to the request.

Same request and response shape — change one base URL and you're live. You gain failover, unified billing across providers, lower latency via our edge, and a single key for every model.

Ready to supercharge your AI stack?

Join thousands of developers building on OmniRouter's stable, professional infrastructure. One key. Every model. Zero friction.