Global AI API Gateway

One OpenAI-compatible API for global AI models.

AITransX is an OpenAI-compatible AI API gateway. Use one base_url and API key across global and China models, with usage and cost in one place.

Get trial API key Try models online

OpenAI SDK compatible $1 trial credit Global + China AI model access

AITransX AI API gateway SDK compatible

Use this as your OpenAI SDK base_url https://api.aitransx.com/v1

01 Authenticate

One API key.

02 Route

Healthy upstreams.

03 Observe

Status and cost.

OpenAI Claude Gemini DeepSeek Qwen Moonshot

Grok Zhipu Volcengine Wenxin Spark Minimax

Service status Current service health

Model catalog Copy exact model IDs

Product

Manage model routes, failover, and usage in one place.

The figures below illustrate the features. For current health, see Service status

gateway / example

Routing

One endpoint for both model markets

Keep the OpenAI request shape and switch only the model parameter.

AITransX gateway POST /v1/chat/completions 200 OK

client OpenAI SDK model: "gpt-4"

AI AITransX /v1

OpenAI Claude Gemini DeepSeek

auth 12ms route 31ms upstream 269ms total 312ms

Reliability

Failover without SDK rewrites

If an upstream fails, try a healthy route without changing application code.

Billing

Usage you can inspect

See tokens, cost, and request status by key, model, and provider.

Security

Prompt bodies are not stored

Forwarded request bodies are not saved as dashboard logs; logs retain status and usage metadata. Optional share links save the content you choose to share.

Common questions

A few things to know before you connect.

Direct answers about the service, compatibility, and billing.

Read the AI API integration guide

What is an AI API gateway?

An AI API gateway receives model requests from your application and forwards them to the appropriate provider. AITransX focuses on model access, with model selection, routing, and usage management.

Can I connect my existing tools and projects?

OpenAI-compatible tools that let you set an endpoint and API key can generally connect. Supported APIs, parameters, and multimodal features depend on the selected model and provider. Check the model catalog and integration docs for details.

Browse the current model catalog

How is model usage billed?

Chat is generally billed by input and output tokens; speech and images may use other units. See the pricing page for rates. Available models and features depend on your account permissions.

Check live pricing

Can I try models without writing code?

Yes. Sign in to try chat and image generation online, check the results, and decide how to use them in your application. Available models and features depend on your account permissions.

Start

Choose how to use AITransX.

For API developers

Connect your app

Keep the OpenAI SDK; set base_url and choose a model ID.

Get trial API key

No code required

Try models online before you integrate

Chat, create images, or validate a model before connecting the API.

Open model workspace