Introduction

Welcome to Puyun AI — a unified AI API gateway that lets you access all major large language models with a single API Key.

What is Puyun AI?

Puyun AI is an AI API aggregation gateway that unifies 40+ upstream AI providers — including OpenAI, Claude, Gemini, Azure, and AWS Bedrock — behind a single API interface. You no longer need to register separate accounts or manage multiple API Keys for each provider. Simply use Puyun AI to call any model.

Core Features

Unified API: Compatible with OpenAI, Anthropic, Google Gemini, and other protocols — access all models through a single endpoint
Automatic Protocol Conversion: When you connect via Claude Code using the Anthropic protocol, the platform automatically adapts requests and responses to the target model's native protocol — no manual adaptation required
Multiple Service Tiers: Three service tiers — Test, Flex, and Enterprise — to meet different needs from testing to enterprise-grade production
Intelligent Routing: Automatic failover and load balancing to ensure high availability
Pay-as-you-go: Only pay for the tokens you actually use, with transparent pricing

Service Tiers

Puyun AI offers three service tiers for different use cases:

Tier	Use Case	Features
Test	Development & testing	Lowest cost, ideal for API debugging and feature verification
Flex	Individual developers	Balanced cost-performance, ideal for personal projects and daily development
Enterprise	Enterprise production	Most stable service with SLA guarantees, ideal for enterprise applications

Model pricing varies across service tiers. See the Usage page for details.

Authentication

All API requests require API Key authentication. Two methods are supported:

Authorization: Bearer <token> — Standard Bearer Token
x-api-key: <your-api-key> — Custom request header

Base URL

The base URL for all API requests is:

https://ai.tracup.com/

Next Steps

Check out the Quick Start guide to make your first API call in minutes.