Introduction

ZeroToken is a unified AI model API gateway. With one API key and one OpenAI-compatible interface you can call global models such as Claude, GPT, Gemini and Grok, as well as Chinese models such as DeepSeek, Qwen, Kimi, Zhipu GLM and MiniMax.

Why ZeroToken

One interface for every model: keep the same base_url and switch models by changing the model parameter. No separate sign-ups, top-ups or SDKs per vendor. Works with the protocols you already use: OpenAI Chat Completions, OpenAI Responses, Anthropic Messages and the native Gemini API are all supported, so existing code and tools usually only need a new URL and key. High availability: a model can be served by several upstream accounts. If one is rate-limited or failing, requests are retried on another automatically. Transparent pay-as-you-go pricing: you pay for the tokens you use, at the prices published in the model catalog and on the pricing page. Control: every key can have its own quota, expiry date, IP allowlist and spend limits, and can be disabled at any time.

Core concepts

Next steps

Follow the Quickstart to send your first request in five minutes. Browse the API reference for every endpoint. Already using Claude Code or Codex? Jump to Integrations.