ZRouter.

THE SUPER INTELLIGENCE ROUTER

The right model.
For every request.

ZRouter is a Super Intelligence Router. Your app calls one OpenAI-compatible endpoint; ZRouter reads each request and sends it to the model that fits, from fast and cheap to your strongest, across OpenAI, Anthropic, Gemini, and more. Switch providers any time without changing your code, env vars, model name, or endpoint.

Open your workspace Explore pricing
MODEL PROVIDERSDEVELOPER TOOLS & AGENTS
OpenAIAnthropicGeminiDeepSeekMistralxAI
ROUTING
BYOK
ONE API
USAGE
ZRouter
Claude CodeCodexCursorOpenCodeClineYour app / SDK
Your tool sends a request through ZRouter to the selected provider. The response comes back through ZRouter.Full coding-agent inference currently requires standalone gateway mode. Check tool setup and account support ↗

PROVIDER CONNECTIONS FOR YOUR NEXT BUILD

OpenAIAnthropicGeminiDeepSeekxAIMistral
Claude CodeCodexCursorOpenCodeClinen8n

Available models and providers depend on your workspace configuration.

CORE CAPABILITIES

A simpler connection.
A clearer view of your AI.

Your requests, model routing, keys, usage, and credits come together in one place.

01

A familiar API

Connect your application using OpenAI-compatible requests.

 02

Models that fit your workload

Keep a stable model name while your routing evolves.

 03

Your keys. Your choice.

Use saved provider credentials on the requests you choose.

 04

Controls in your account

Manage API keys, budgets, and request limits.

 05

See what happened

Follow model activity, token usage, and request logs.

 06

Understand every charge

Keep prepaid credit and settled charges in one ledger.
ENDPOINTS & PROTOCOLSONE GATEWAY

Keep your requests familiar.

Use your ZRouter key and a model from your dashboard. The gateway translates requests for the selected provider.

Chat completionsPOST /v1/chat/completions
ResponsesPOST /v1/responses
MessagesPOST /v1/messages
EmbeddingsPOST /v1/embeddings

Endpoint support depends on your selected model. See the API guide for account limitations.

Read the API guide ↗
VIRTUAL MODEL ROUTINGONE MODEL NAME

Change the route, keep the connection.

A virtual model is a reusable name for routing requests to provider models. Choose targets and scheduled routes in your workspace.

Your applicationyour-virtual-model
Primary targetFallback target

Illustrated routing configuration. Targets depend on the models available to your account.

Create a virtual model ↗
BRING YOUR OWN KEYSENCRYPTED STORAGE

Your provider account, connected.

Save supported provider credentials with encryption at rest. Enable BYOK for individual requests with a request header.

Provider credential••••••••••••••••••••••••SAVED IN YOUR ACCOUNT
X-ZRouter-BYOK: true

Your provider bills model usage directly. ZRouter routing fees, if configured, are separate.

Read the BYOK guide ↗
ACCOUNT CONTROLSYOUR WORKSPACE

Give every application its own key.

Create and revoke API keys from your account. Manage budgets and rate-limit rules when enabled in your workspace.

API keysCreate · Revoke
BudgetsTrack estimated spend
Rate limitsManage request rules
Account accessYour personal dashboard

Budget estimates do not replace your prepaid balance or erase settled charges.

Explore account controls ↗
USAGE & VISIBILITYONE DASHBOARD

From a request to a clearer picture.

Review token usage, model activity, and available request logs. Test an available model in the playground before integrating it.

Illustrated usage chart. Your dashboard shows your own account activity.

Explore your workspace ↗
CREDITS & BILLINGPAY BY THE TOKEN

A balance you can follow.

Add prepaid USD credit, share it across your account keys, and review settled charges in your billing ledger.

Minimum credit purchase$5
Billing unitInput & output tokens
Payment flowStripe-hosted Checkout
Charge detailsBilling & credits ledger

Credit is reserved before a request and settled when complete usage is reported.

Understand credits and billing ↗

WORK WHERE YOU BUILD

One account.
Three ways to manage it.

Use the dashboard for a visual view, zctl for repeatable commands, or MCP to work with your AI assistant.

See it in the dashboard

Compare usage, inspect available request logs, manage credits, and test models in the playground. Start here when you want to see the whole account.

Open the quickstart ↗

Automate it with zctl

Create and edit virtual routers, manage keys, and review usage from your terminal. JSON output and router exports make account tasks easier to repeat in scripts.

Install the zctl CLI ↗

Ask your agent through MCP

Connect Claude Code, Cursor, or another compatible MCP client. Ask about usage, investigate errors, and manage routers, API keys, and enabled account limits.

Connect an MCP client ↗

CLI and MCP management access is limited to your account. Connecting MCP does not change your agent's inference provider. See each tool's integration guide for model-request support.

GETTING STARTED

Three steps.
Then build something great.

A workspace, an API key, and your first request. Start with the tools you already use.

YOUR WORKSPACE
API keyzr_••••••••••••••••
Prepaid creditsBring your own key

One account. Your choice.

01

Create an account and key

Sign up, choose prepaid credits or BYOK, and create an API key in your dashboard.

YOUR APPLICATION
client = OpenAI(
  api_key="YOUR_ZROUTER_KEY",
  base_url="https://zrouter.si/v1"
)

Your SDK. A new connection.

02

Connect your application

Set your gateway base URL and ZRouter API key. Pick an available model from your account.

YOUR FIRST REQUEST
POST /v1/chat/completions

“Hello, ZRouter”

Model responseToken usage

Test in your app or the playground.

03

Send your first request

Make a call, check the response, and follow its usage and settled charges in your workspace.

Follow the quickstart ↗
One APIA familiar connection
Your keysBring your own provider
$5Minimum credit purchase
No subscriptionPay by the token

PRICING

Make the most of every token.

Prepaid credits for gateway-funded requests. Your own provider keys when you prefer direct provider billing.

One workspace, no subscription.

Everything you need to connect, test, and follow your usage.

  • Account API keys and a shared credit balance
  • Available models and a personal playground
  • Token usage and a billing ledger
  • Encrypted BYOK credential storage
  • Documentation and integration guides
$5 minimum credit purchase
Create your account ↗

New accounts start at $0. No card needed to sign up.

Model pricing

USD per 1 million tokens.

Model / providerInputOutput
anthropic/claude-fable-5$10$50
anthropic/claude-fable-5-1$10$50
anthropic/claude-haiku-4-5-20251001$1$5
anthropic/claude-opus-4-5-20251101$5$25
anthropic/claude-opus-4-6$5$25
anthropic/claude-opus-4-7$5$25
BYOK routing fees$0$0

Current loaded rates. Context tiers, cache writes, and metered fallback attempts can affect the final charge. BYOK provider charges are separate.

View complete pricing ↗

FAQS

A few things
before you connect.

Practical answers for your first request.

What is ZRouter?

ZRouter is an AI gateway that connects your application to multiple model providers through one OpenAI-compatible API. Your workspace brings API keys, virtual models, usage, provider credentials, and prepaid credits together. Manage the same account in the dashboard, through the zctl CLI, or with an AI assistant using MCP.

Who is ZRouter for?

Developers and teams building AI features who want a consistent API, the flexibility to choose models, and a clear view of usage and charges.

Can I manage ZRouter from my terminal or AI assistant?

Yes. Use the zctl CLI for account commands and scripts. Connect an assistant through the account MCP server to inspect usage and manage routers, API keys, and enabled limits. MCP management does not change your assistant's inference provider.

Can I keep using my existing SDK?

Use an OpenAI-compatible SDK with your ZRouter base URL and account API key. Endpoint and parameter support depend on the selected model. Read the API guide for supported requests and account limitations.

What is a virtual model?

A virtual model is a stable name that routes to configured provider models. It lets you change targets or schedule routes without changing the model name in your application. Read the virtual models guide for who they help, how to create one, and available routing options.

Can I bring my own provider keys?

Yes. Save supported provider credentials in your account, then enable BYOK on a request. Your provider bills its usage directly. ZRouter routing fees, if configured, are charged separately.

Is there a subscription or free credit?

There is no monthly subscription. New accounts start at $0. Add at least $5 per credit purchase, or connect your own supported provider key. BYOK routing fees, if any, also require a ZRouter balance.

Where can I see my usage and charges?

Your dashboard shows token usage, model activity, and available request logs. Billing & credits contains your balance and settled charge ledger. Usage estimates can differ from final charges.

YOUR NEXT REQUEST STARTS HERE

Build with AI.
Keep the controls.

One account for model access, routing, and usage. Manage it from your dashboard, terminal, or AI assistant.

Create your account Read the quickstart No card needed to create an account.