Get API Key

One API key for every major AI model

An OpenAI-compatible endpoint that routes to OpenAI, Anthropic, Google, DeepSeek, Qwen and more — with unified billing and per-request logs.

Get an API key

One key, many models

OpenAI, Anthropic, Google, DeepSeek and more behind a single endpoint.

Drop-in compatible

OpenAI-compatible API — change the base URL, keep your code.

Predictable billing

Token-level metering with per-request logs you can reconcile.

Automatic failover

Unhealthy channels leave the pool before your users notice.

Live in three steps

Step 1

Create a key

Sign up, top up a balance, and generate an API key in the console.

Step 2

Change the base URL

Point your OpenAI SDK at the gateway. Your existing request code stays the same.

Step 3

Ship and monitor

Watch token usage and spend per key, and switch models by changing one string.

Why route through a gateway

One integration instead of one per provider.

Direct integrations

  • ✕A separate account, key and invoice per provider
  • ✕Your code changes when you switch models
  • ✕One provider outage is a visible error
  • ✕Usage is spread across several dashboards

Through TokWork

  • ✓One key and one balance for every provider
  • ✓Swap models by changing a single string
  • ✓Failing channels leave the routing pool automatically
  • ✓Token-level usage and cost in one log

Built for people shipping AI features

TokWork is the gateway layer between your application and the model providers. Routing, retries, metering and logging live in one place, so your code talks to one endpoint and your team reconciles one bill.

1

Endpoint for every provider

OpenAI

Compatible request format

Per request

Usage and cost logging

24/7

Automated channel health checks

Read the docs

Frequently asked questions

The things people ask before their first request.

What is TokWork?
TokWork is an API gateway for large language models. You get one API key and one OpenAI-compatible endpoint, and TokWork routes each request to the provider of your choice.
Do I need to change my existing code?
No. The endpoint is OpenAI-compatible, so in most SDKs you only change the base URL and the API key. Requests in the Anthropic and Gemini formats are accepted as well.
Which models are available?
Models from OpenAI, Anthropic, Google, DeepSeek, Qwen and other providers are routed through the same key. The full list with live pricing is returned by the /v1/models endpoint.
How is usage billed?
Every request is metered by input and output tokens and deducted from your account balance. Per-request logs are available in the console, so you can reconcile spend against your application.
What happens if a provider is unavailable?
Channels are health-checked continuously and removed from the routing pool when they fail, so a request is served by a working channel instead of returning an error.
How do I get started?
Create an account, top up a balance, generate an API key, and point your SDK at the gateway base URL. The quickstart in the docs is a five-line change.

Start with one key

Create an account, top up, and send your first request in a few minutes.

Get an API key