Introduction

NemoRouter is an enterprise LLM gateway that gives your team one API key, one bill, and total control. Call any LLM through a single OpenAI-compatible endpoint, with budget enforcement, guardrails, and smart routing enforced in the request path — not reported after the fact. 240+ models are live today across six provider clouds — Alibaba US, OpenAI, AWS Bedrock, Azure Foundry, Google Vertex AI, and Anthropic — with more landing on a rolling basis.

Why NemoRouter?

Managing multiple LLM providers is painful. Different API keys, different billing dashboards, different SDKs, different rate limits. NemoRouter eliminates all of that.

Key Features

One Key, One Bill

Buy credits, call any model. NemoRouter manages all provider keys behind the scenes. You never touch a provider API key.

OpenAI-Compatible API

Use any OpenAI SDK or HTTP client. Just change the base URL to https://api.nemorouter.ai/v1 and swap your API key. Your existing code works immediately.

One Endpoint, Growing Catalog

Access 240+ models through a single endpoint — no per-provider keys or SDKs:

• Alibaba US (live) — Qwen and DeepSeek models
• OpenAI (live) — GPT and o-series models
• AWS Bedrock (live) — Claude, Llama, DeepSeek, Nova, and Qwen
• Azure Foundry (live) — OpenAI-format and native/partner models
• Google Vertex AI (live) — Gemini models
• Anthropic (live) — Claude Opus, Sonnet, and Haiku

See the Models page for the current live catalog.

Enterprise Controls

• Guardrails — Content safety, PII detection, prompt injection prevention
• Budget controls — Spending limits per key, team, or organization
• Rate limits — RPM and TPM controls to prevent runaway costs
• Team management — Roles, permissions, and per-team budgets
• Analytics — Real-time spend tracking, usage breakdowns, and request logs

Getting Started

Getting up and running with NemoRouter takes three steps:

1. Create an Account

Sign up at nemorouter.ai and complete the onboarding flow. You'll set up your organization and choose a plan.

2. Get Your API Key

After onboarding, create an API key from the API Keys page in your dashboard. Your key will look like sk-nemo-xxxx. Copy it immediately — it's only shown once.

3. Make Your First Request

curl https://api.nemorouter.ai/v1/chat/completions \
  -H "Authorization: Bearer $NEMOROUTER_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gemini-2.5-flash",
    "messages": [
      {"role": "user", "content": "Hello, NemoRouter!"}
    ]
  }'

That's it. You're calling Gemini 2.5 Flash through NemoRouter. Change the model field to gemini-2.5-pro or any other model in the live catalog — no configuration needed.

How It Works

Your App → NemoRouter API → Provider (OpenAI, Anthropic, Google, etc.)

NemoRouter sits between your application and the LLM providers. Every request flows through NemoRouter, which handles:

1. Authentication — Validates your API key and checks permissions
2. Guardrails — Applies any configured safety controls
3. Routing — Sends the request to the right provider
4. Cost tracking — Records token usage and spend against your credit balance
5. Response — Returns the provider's response with NemoRouter headers for observability

Pricing

NemoRouter uses a credit-based system with a simple, honest platform fee:

Every feature is available on every plan — guardrails, budgets, routing, prompts, analytics, SSO, audit logs. Plans change only the platform fee and guaranteed throughput, never the feature set. The platform fee is charged on top of your credit purchases, never deducted. Load your first $5 in credits and we add a $10 bonus — start with $15 for the full catalog.

Next Steps

• Quick Start — Make your first API call in under 2 minutes
• Authentication — Learn how API keys and auth work
• Chat Completions — Explore the Chat Completions API
• Python SDK — Get started with the Python SDK